TR-808 Bass Drum has variable punch & decay

classic Classic list List threaded Threaded
15 messages Options
Reply | Threaded
Open this post in threaded view
|

TR-808 Bass Drum has variable punch & decay

Justin Morgan
Hey all...
Just got a TR-808 again.
Overall, it's in very good condition.  Dirty and dusty. Could use a good cleaning inside and out.
I'm going through the sounds, and checking for problems.

The only issue is the the beloved BD.
It seems the overall drive (punch/loudness) and the decay wants to fluctuate. (ie, it's not constant)
This is mainly noticed when I have the decay level set to max (all the way to the right)
It happens with and without the added Accent.

Is this a sign of a major problem?
OR...  Would a routine cleaning and tune up fix this?

Honestly, my skill set at fixing stuff is poor.
Any real work will have to be done by someone else.
Just trying to decide if it's "broken" and needs repair/parts... or just a deep clean/tune.

Seriously, any help or advice would be great.
I'd love to get this 808 back to 100%. 
Which reminds me, if anyone has an extra 16 step key set... I'll buy it.
Seems like Cy doesn't have them on ebay... right now.

Thanks tons!
Justin

Reply | Threaded
Open this post in threaded view
|

Re: TR-808 Bass Drum has variable punch & decay

Pete-54
 Your decay pot might be bad. I would open it up and try to clean that pot really well. If that doesn't work I would replace that pot. Another thing that might be bad is a capacitor in the kick circuit. Look for any electrolytic capacitors and replace them. I highly doubt there are any tantalum caps, but always replace those with an electrolytic cap. I'm not sure if there is any electrolytic caps in the kick circuit, but it doesn't hurt to look for them. My 808 doesn't have that problem with the kick. Good luck.

pete



On Tue, Mar 8, 2011 at 10:36 AM, Justin Morgan <[hidden email]> wrote:
Hey all...
Just got a TR-808 again.
Overall, it's in very good condition.  Dirty and dusty. Could use a good cleaning inside and out.
I'm going through the sounds, and checking for problems.

The only issue is the the beloved BD.
It seems the overall drive (punch/loudness) and the decay wants to fluctuate. (ie, it's not constant)
This is mainly noticed when I have the decay level set to max (all the way to the right)
It happens with and without the added Accent.

Is this a sign of a major problem?
OR...  Would a routine cleaning and tune up fix this?

Honestly, my skill set at fixing stuff is poor.
Any real work will have to be done by someone else.
Just trying to decide if it's "broken" and needs repair/parts... or just a deep clean/tune.

Seriously, any help or advice would be great.
I'd love to get this 808 back to 100%. 
Which reminds me, if anyone has an extra 16 step key set... I'll buy it.
Seems like Cy doesn't have them on ebay... right now.

Thanks tons!
Justin




--
www.p23.biz
Reply | Threaded
Open this post in threaded view
|

RE: TR-808 Bass Drum has variable punch & decay

el macaco
If the kick is still decaying when it receives the next trigger you will get inconsistant / different sounding kicks, if the pattern you are running displays this problem try reducing the decay or lowering the tempo and you should hear the 'problem' go away.  This is a common misunderstanding with people who use the 808 for the first time.  Then again, it could be something wrong if it happens without the kick being retriggered in the decay stage.
 

 

Date: Tue, 8 Mar 2011 11:23:05 -0800
From: [hidden email]
To: [hidden email]
Subject: Re: [AH] TR-808 Bass Drum has variable punch & decay

 Your decay pot might be bad. I would open it up and try to clean that pot really well. If that doesn't work I would replace that pot. Another thing that might be bad is a capacitor in the kick circuit. Look for any electrolytic capacitors and replace them. I highly doubt there are any tantalum caps, but always replace those with an electrolytic cap. I'm not sure if there is any electrolytic caps in the kick circuit, but it doesn't hurt to look for them. My 808 doesn't have that problem with the kick. Good luck.

pete



On Tue, Mar 8, 2011 at 10:36 AM, Justin Morgan <[hidden email]> wrote:
Hey all...
Just got a TR-808 again.
Overall, it's in very good condition.  Dirty and dusty. Could use a good cleaning inside and out.
I'm going through the sounds, and checking for problems.

The only issue is the the beloved BD.
It seems the overall drive (punch/loudness) and the decay wants to fluctuate. (ie, it's not constant)
This is mainly noticed when I have the decay level set to max (all the way to the right)
It happens with and without the added Accent.

Is this a sign of a major problem?
OR...  Would a routine cleaning and tune up fix this?

Honestly, my skill set at fixing stuff is poor.
Any real work will have to be done by someone else.
Just trying to decide if it's "broken" and needs repair/parts... or just a deep clean/tune.

Seriously, any help or advice would be great.
I'd love to get this 808 back to 100%. 
Which reminds me, if anyone has an extra 16 step key set... I'll buy it.
Seems like Cy doesn't have them on ebay... right now.

Thanks tons!
Justin




--
www.p23.biz
Reply | Threaded
Open this post in threaded view
|

Re: TR-808 Bass Drum has variable punch & decay

Justin Morgan
In reply to this post by Justin Morgan
Thanks for the tips on cleaning.
I'll do best to get the grim and dust removed... (with air)
I have contact cleaner too, but will probably only use in small amounts.  (or not at all)
If pots and/or caps need replacing, I'll consider my options for fixing...  thanks again.

As for the retriggered decay... yeah, I know what ya mean.
This isn't the case.  This problem happens even if I program the first step only, out of 16... and it loops
with the variable decay/punch.
But good reply... as I'm sure some have been in that situation before.
Reply | Threaded
Open this post in threaded view
|

Re: TR-808 Bass Drum has variable punch & decay

Julian-71
In reply to this post by Justin Morgan
Do be aware that individual 808s vary considerably from one another.  The bass drum is the most discussed, but the other voices also differ significantly.
 
There are things you can do to meddle with the voices, to adjust them to your liking, but your new machine may actually be equally as 'valid' as your old one, but just different : )
 
The bass drum will not sound the same on every step either.  Its a common cause of concern with the 808 - i would have thought your old one also perfomed similarly?
 
Julian
 
----- Original Message -----
Sent: Tuesday, March 08, 2011 6:36 PM
Subject: [AH] TR-808 Bass Drum has variable punch & decay

Hey all...
Just got a TR-808 again.
Overall, it's in very good condition.  Dirty and dusty. Could use a good cleaning inside and out.
I'm going through the sounds, and checking for problems.

The only issue is the the beloved BD.
It seems the overall drive (punch/loudness) and the decay wants to fluctuate. (ie, it's not constant)
This is mainly noticed when I have the decay level set to max (all the way to the right)
It happens with and without the added Accent.

Is this a sign of a major problem?
OR...  Would a routine cleaning and tune up fix this?

Honestly, my skill set at fixing stuff is poor.
Any real work will have to be done by someone else.
Just trying to decide if it's "broken" and needs repair/parts... or just a deep clean/tune.

Seriously, any help or advice would be great.
I'd love to get this 808 back to 100%. 
Which reminds me, if anyone has an extra 16 step key set... I'll buy it.
Seems like Cy doesn't have them on ebay... right now.

Thanks tons!
Justin

Reply | Threaded
Open this post in threaded view
|

Re: TR-808 Bass Drum has variable punch & decay

Justin Morgan
Thanks Julian...

You are correct.   But this is more than just subtle analog differences in tone/feel.
I've owned or used 3 or 4 TR-808s before... and this is the first time I've noticed these
larger differences from step to step.
After cleaning tonight, I'll try to record a 30 sec wav for upload...
Just so everyone will hear what I'm hearing.  :-)
Thanks for your reply... Much appreciated.

Justin



On Tue, Mar 8, 2011 at 2:50 PM, Julian <[hidden email]> wrote:
Do be aware that individual 808s vary considerably from one another.  The bass drum is the most discussed, but the other voices also differ significantly.
 
There are things you can do to meddle with the voices, to adjust them to your liking, but your new machine may actually be equally as 'valid' as your old one, but just different : )
 
The bass drum will not sound the same on every step either.  Its a common cause of concern with the 808 - i would have thought your old one also perfomed similarly?
 
Julian
 
----- Original Message -----
Sent: Tuesday, March 08, 2011 6:36 PM
Subject: [AH] TR-808 Bass Drum has variable punch & decay

Hey all...
Just got a TR-808 again.
Overall, it's in very good condition.  Dirty and dusty. Could use a good cleaning inside and out.
I'm going through the sounds, and checking for problems.

The only issue is the the beloved BD.
It seems the overall drive (punch/loudness) and the decay wants to fluctuate. (ie, it's not constant)
This is mainly noticed when I have the decay level set to max (all the way to the right)
It happens with and without the added Accent.

Is this a sign of a major problem?
OR...  Would a routine cleaning and tune up fix this?

Honestly, my skill set at fixing stuff is poor.
Any real work will have to be done by someone else.
Just trying to decide if it's "broken" and needs repair/parts... or just a deep clean/tune.

Seriously, any help or advice would be great.
I'd love to get this 808 back to 100%. 
Which reminds me, if anyone has an extra 16 step key set... I'll buy it.
Seems like Cy doesn't have them on ebay... right now.

Thanks tons!
Justin


Reply | Threaded
Open this post in threaded view
|

Re: TR-808 Bass Drum has variable punch & decay

Justin Morgan
Hey guys...

Just a follow up.
Here's a short snippet of the 808 BD:


It's one pattern, looped...
I'm not messing with any 808 knobs while recording.
BD knobs: Level 100%, Tone 50%, Decay 100%

You can hear the variation.  It's all over the place.

Just sprayed air into knob area.
Prob will take apart tomorrow.
But, I think it might be a malfunctioning part.





On Tue, Mar 8, 2011 at 2:57 PM, Justin Morgan <[hidden email]> wrote:
Thanks Julian...

You are correct.   But this is more than just subtle analog differences in tone/feel.
I've owned or used 3 or 4 TR-808s before... and this is the first time I've noticed these
larger differences from step to step.
After cleaning tonight, I'll try to record a 30 sec wav for upload...
Just so everyone will hear what I'm hearing.  :-)
Thanks for your reply... Much appreciated.

Justin




On Tue, Mar 8, 2011 at 2:50 PM, Julian <[hidden email]> wrote:
Do be aware that individual 808s vary considerably from one another.  The bass drum is the most discussed, but the other voices also differ significantly.
 
There are things you can do to meddle with the voices, to adjust them to your liking, but your new machine may actually be equally as 'valid' as your old one, but just different : )
 
The bass drum will not sound the same on every step either.  Its a common cause of concern with the 808 - i would have thought your old one also perfomed similarly?
 
Julian
 
----- Original Message -----
Sent: Tuesday, March 08, 2011 6:36 PM
Subject: [AH] TR-808 Bass Drum has variable punch & decay

Hey all...
Just got a TR-808 again.
Overall, it's in very good condition.  Dirty and dusty. Could use a good cleaning inside and out.
I'm going through the sounds, and checking for problems.

The only issue is the the beloved BD.
It seems the overall drive (punch/loudness) and the decay wants to fluctuate. (ie, it's not constant)
This is mainly noticed when I have the decay level set to max (all the way to the right)
It happens with and without the added Accent.

Is this a sign of a major problem?
OR...  Would a routine cleaning and tune up fix this?

Honestly, my skill set at fixing stuff is poor.
Any real work will have to be done by someone else.
Just trying to decide if it's "broken" and needs repair/parts... or just a deep clean/tune.

Seriously, any help or advice would be great.
I'd love to get this 808 back to 100%. 
Which reminds me, if anyone has an extra 16 step key set... I'll buy it.
Seems like Cy doesn't have them on ebay... right now.

Thanks tons!
Justin



Reply | Threaded
Open this post in threaded view
|

Re: TR-808 Bass Drum has variable punch & decay

Pete-54
 Yea that's not good. Something is wrong in there for sure. I just checked the BD schematic, and there is an electrolytic cap in the circuit that is connected to the decay pot. I would replace that cap. That's the cap that sets the decay. I would guess that it's bad and it's tolerance is fluctuating. These electro caps are 20 years old! The liquid inside has prob dried out. Old caps also produce noise. It is a very common problem with old gear. You can try replacing it and then see if that works. Unfortunately dissembling old gear can be tricky, so be careful you don't damage something cuz it's hard to find replacements.


pete




On Tue, Mar 8, 2011 at 3:20 PM, Justin Morgan <[hidden email]> wrote:
Hey guys...

Just a follow up.
Here's a short snippet of the 808 BD:


It's one pattern, looped...
I'm not messing with any 808 knobs while recording.
BD knobs: Level 100%, Tone 50%, Decay 100%

You can hear the variation.  It's all over the place.

Just sprayed air into knob area.
Prob will take apart tomorrow.
But, I think it might be a malfunctioning part.





On Tue, Mar 8, 2011 at 2:57 PM, Justin Morgan <[hidden email]> wrote:
Thanks Julian...

You are correct.   But this is more than just subtle analog differences in tone/feel.
I've owned or used 3 or 4 TR-808s before... and this is the first time I've noticed these
larger differences from step to step.
After cleaning tonight, I'll try to record a 30 sec wav for upload...
Just so everyone will hear what I'm hearing.  :-)
Thanks for your reply... Much appreciated.

Justin




On Tue, Mar 8, 2011 at 2:50 PM, Julian <[hidden email]> wrote:
Do be aware that individual 808s vary considerably from one another.  The bass drum is the most discussed, but the other voices also differ significantly.
 
There are things you can do to meddle with the voices, to adjust them to your liking, but your new machine may actually be equally as 'valid' as your old one, but just different : )
 
The bass drum will not sound the same on every step either.  Its a common cause of concern with the 808 - i would have thought your old one also perfomed similarly?
 
Julian
 
----- Original Message -----
Sent: Tuesday, March 08, 2011 6:36 PM
Subject: [AH] TR-808 Bass Drum has variable punch & decay

Hey all...
Just got a TR-808 again.
Overall, it's in very good condition.  Dirty and dusty. Could use a good cleaning inside and out.
I'm going through the sounds, and checking for problems.

The only issue is the the beloved BD.
It seems the overall drive (punch/loudness) and the decay wants to fluctuate. (ie, it's not constant)
This is mainly noticed when I have the decay level set to max (all the way to the right)
It happens with and without the added Accent.

Is this a sign of a major problem?
OR...  Would a routine cleaning and tune up fix this?

Honestly, my skill set at fixing stuff is poor.
Any real work will have to be done by someone else.
Just trying to decide if it's "broken" and needs repair/parts... or just a deep clean/tune.

Seriously, any help or advice would be great.
I'd love to get this 808 back to 100%. 
Which reminds me, if anyone has an extra 16 step key set... I'll buy it.
Seems like Cy doesn't have them on ebay... right now.

Thanks tons!
Justin






--
www.p23.biz
Reply | Threaded
Open this post in threaded view
|

Re: TR-808 Bass Drum has variable punch & decay

Robin Whittle
Short version:  1 - Modifying the TR-808 kick (BD) circuit to give any
                     decay time, including to self-resonance.

                 2 - Why the BD circuit gives erratic sounds when
                     triggered while it is still resonating.

                 3 - Altering the tuning of the BD circuit.

Hi Justin, Pete and others,

In regard to the Justin's recording:

   http://soundcloud.com/joustin/weird-808-bd

Pete wrote, in part:

>   Yea that's not good. Something is wrong in there for sure.

I disagree entirely.

The maximum decay time of a TR-808 BD circuit depends on the gain of a
feedback circuit, which somewhat resembles a Twin-T resonator / (almost)
oscillator circuit.  The exact gain depends on ratios of resistors.
Since resistor values vary slightly from one machine to another, the
gain will vary somewhat from one machine to the next.  To make up some
figures, if the gain was 0.995 we might expect a decay time of X
milliseconds.  If the gain was 0.997, the decay time might be nearly
twice X milliseconds.  So fractional percentage variations in the
resistor values can lead to highly audible differences in the maximum
decay time.

An excellent copy of the TR-808 service manual can be found at:

   http://www.synfo.nl/

Page 9 shows the Twin-T circuit:

   http://en.wikipedia.org/wiki/RC_oscillator

The -C-+-C-
        |
        R
        |

part of the feedback circuit is C41, (R166 + R165) and C42.  R165 is
shorted out during and just after the trigger pulse (which is ~1msec) by
Q43, to give a much higher resonant pitch, which is part of the "click"
sound at the start.

The R-C-R part of the Twin-T circuit does not appear in the TR-808 BD.
Instead, we have R167 (1M) going from the IC12 pin 1 output to its
negative input (pin 2) plus an inverted and somewhat less than 1.0 gain
second amplifier (IC12 pins 5, 6 and 7) feeding into the middle of the
C-R-C circuit, via R170 (470k).  This second amplifier, with a negative
polarity output, driving the middle of the C-R-C network which itself is
strung across the inverting amplifier of the first amplifier IC12 pins
2, 3 and 1, means we get positive feedback, but not quite to the gain of
1.0, which would sustain continual oscillation.

When VR6 is CW (clockwise), the gain of the inverting amplifier is
around -0.9, because the input resistor is 47k and the feedback resistor
is (in AC terms) 47k in parallel with 500k, which is about 0.9 of 47k.
The high value (33uF) of C43 in series with the 500k CW value of VR6
would have (according to my rough calculations) a 1/ (2 pi R C)
frequency of about 10Hz, so the gain of this amplifier might be a little
higher at the BD frequency.

Reducing VR6 by turning it ACW (anti-clockwise) would reduce the gain of
this amplifier which reduces (VR6 ACW = 0 ohms) the resonance of the
whole circuit, down to zero, or close to zero, since it becomes R164
driving into R169 in parallel with C43.  So this amplifier system's gain
can be varied from about -0.90 (VR6 CW) to about -0.000..? (VR6 ACW).

The resonance when VR6 is CW would depend on the value of these resistors:

   R167  R164  VR6  R169  R170

It may also be affected by R161 and C39, going back through R160.

This circuit is rather complex and I wouldn't want to have to write a
definitive treatise on exactly what it does.  I doubt the designers did
this - I guess they mucked around until they got the sound they wanted,
which I think is a perfectly good approach.

Without doing the complete qualitative and quantitative analysis, we can
simply tweak the circuit for more gain in the second amplifier, so we
get more feedback, which with the phase shifts due to the capacitors,
and the inverting nature of the first amplifier, leads to a higher
resonance, longer resonant notes and potentially to self-resonance on a
continual basis without any trigger from the sequencer.

I wrote about this to Analogue Heaven on 11 November 1996.  I suggested
soldering a 390k resistor in parallel with R164 (47k).

This means that the gain of the amplifier with VR6 CW is:

    (47k || 500k)   /  (47k || 390k)

where "||" means "in parallel with".

This is:

     (1 / ( (1 / 47000)  + ( 1 / 500000) )
   / (1 / ( (1 / 47000)  + ( 1 / 390000) )

  ~=   42961  / 41956  ~=  1.024

This should enable continual self-oscillation when VR6 is CW.  If not,
try a 330k.

Now, by backing of VR6 from the fully CW position, you can get resonance
as long as you like, to the point of self-resonance without any trigger
pulse.

The exact amount you would need to turn it to get a given resonance
would depend on the values of all these resistors.  With this mod, if
you are only interested in the BD not self-resonating, then you can get
any decay time you like - thereby eliminating all this stuff about "one
TR-808 BD circuit sounding different from another".


Now, to the second question of why the BD gives "erratic" responses as
demonstrated in Justin's sound file.

We can imagine the BD circuit being a sine-wave almost resonant
oscillator, which is altered in three ways when a trigger pulse arrives.

   1 - The C40, R163, R162 and D53 circuit provides a positive pulse
       to the BD's resonator amplifier's positive input (IC12 pin 1).

       This, in the absence of any other stimulus or current oscillation
       will cause the circuit to oscillate, starting with a positive
       sine-like cycle starting at the time of the trigger pulse.

   2 - Q41 shorts out C38 which means Q42 turns off and remains off until
       R156 charges up C38 enough, after the end of the 1msec trigger
       pulse, to turn Q42 on again.   During this time, Q43 is off, so
       R165 is shorted out, giving a higher resonant frequency.

   3 - There's a very small (0.6 volts, due to D52 turning on) kick to
       the system via C39 and R161 when Q42 turns on, and a big kick in
       the opposite direction when it turns off (which depends on R156,
       C38, the gain of Q42, the temperature of Q42 and the precise
       value of Jupiter's gravitational influence on all these
       components).

Overall, we can think of this as a temporary switch of resonance
frequency to a much higher frequency plus two sources of energy which
drive the system to ring afterwards - to create a sinusoidal signal of
decreasing amplitude, just like the resonance of a simple drum.

If the system is not resonating at all - it has no sinusoidal change
going on (that is, it is silent) - then when we get a trigger pulse of a
given amplitude (5V normally, but up to about 14V for an accented beat)
then we will get a completely predictable ringing waveform thereafter.
The amplitude of this ringing waveform will always be the same from one
beat to the next - *provided* we allow enough time between the trigger
pulses for the resonance to decay to "zero".  In fact, it never decays
to absolute zero - but if we let it decay so far that we can't hear it,
which means it is, say, 1/1000 the amplitude it was at the start of the
note.

*However* if, when the circuit is still resonating, we give it another
trigger pulse, then what will happen?

It depends on:

   1 - The amplitude of the trigger pulse.

   2 - The exact part of the cycle the circuit is at in its nearly
       self-resonant decay.

For the sake of simplicity, I will ignore the frequency change which
follows the trigger pulse.  We can imagine the trigger pulse being a
single, simple, thing which puts "positive" energy into the nearly
self-resonant circuit, of let's say "1 volt".  That is: if the trigger
pulse arrived when there was no self-oscillation, then the system would
ring with a 1V peak (2V peak to peak) sine-wave which then decays at
some rate.

Now let's consider two situations when the trigger pulse arrives when
the circuit is still resonating from a previous trigger pulse:

   A: The trigger pulse arrives when the circuit is at a +0.3V peak.

   B: The trigger pulse arrives when the circuit is at a -0.3V peak.

Broadly speaking, what will result, is a new "note", with the following
characteristics:

   A: 1.3 volt peak.  This is louder than the first note.

   B: 0.7 volt peak.  This is quieter than the first note.


So the variations in volume, softer and louder, which appear in Justin's
recording are due to the trigger pulse arriving at various times in the
oscillation.  The trigger pulse is always the same (except of course if
accented notes are used, which I guess is the case in this recording)
but in the B case, the trigger pulse's energy went into quenching the
oscillation, whereas in the A case, it added to the existing energy, and
resulted in a louder note than usual.

Listening to Justin's recording, my impression is that any TR-808 would
do this if it had the same sequence of trigger pulses and the same
self-resonance.  The exact details of which notes were louder and
software would depend on the precise timing of the trigger pulses, the
precise state of the oscillation cycle at that time - and so the precise
values of resistors, the temperature of the components etc.


As I wrote in 1996, the tuning of the BD circuit can be changed by
replacing R165 with a 100k pot.  ACW means higher pitches.  CW means
lower pitches.  A 100k log pot would be better than a linear one, but
either would do the job.  This involves cutting a track on the top side
of the board.  There's no easy way of mounting a pot there, but
sufficiently small pots can be mounted in the side panels.

Only technicians should work on TR-808s.  There is the danger of static
electricity damage to the circuitry, such as zapping the CPU or a memory
chip.  Also, there are exposed mains wires inside the machine, so anyone
working on the machine while it is powered up risks DEATH.


  - Robin    http://www.firstpr.com.au/rwi/dfish/
             http://www.firstpr.com.au/rwi/tr-808/

Reply | Threaded
Open this post in threaded view
|

Re: TR-808 Bass Drum has variable punch & decay

Greg Morgan-4
So you are saying that this particular 808 came off the factory line with this issue or is this something that happens to random 808's as the circuits age?

I apologize for asking uneducated questions, just trying to follow how the problem came to be in the first place.

Greg

On Tue, Mar 8, 2011 at 10:26 PM, Robin Whittle <[hidden email]> wrote:
Short version:  1 - Modifying the TR-808 kick (BD) circuit to give any
                   decay time, including to self-resonance.

               2 - Why the BD circuit gives erratic sounds when
                   triggered while it is still resonating.

               3 - Altering the tuning of the BD circuit.

Hi Justin, Pete and others,

In regard to the Justin's recording: Pete wrote, in part:


 Yea that's not good. Something is wrong in there for sure.

I disagree entirely.

The maximum decay time of a TR-808 BD circuit depends on the gain of a feedback circuit, which somewhat resembles a Twin-T resonator / (almost) oscillator circuit.  The exact gain depends on ratios of resistors. Since resistor values vary slightly from one machine to another, the gain will vary somewhat from one machine to the next.  To make up some figures, if the gain was 0.995 we might expect a decay time of X milliseconds.  If the gain was 0.997, the decay time might be nearly twice X milliseconds.  So fractional percentage variations in the resistor values can lead to highly audible differences in the maximum decay time.

An excellent copy of the TR-808 service manual can be found at:

 http://www.synfo.nl/

Page 9 shows the Twin-T circuit:

 http://en.wikipedia.org/wiki/RC_oscillator

The -C-+-C-
      |
      R
      |

part of the feedback circuit is C41, (R166 + R165) and C42.  R165 is shorted out during and just after the trigger pulse (which is ~1msec) by Q43, to give a much higher resonant pitch, which is part of the "click" sound at the start.

The R-C-R part of the Twin-T circuit does not appear in the TR-808 BD. Instead, we have R167 (1M) going from the IC12 pin 1 output to its negative input (pin 2) plus an inverted and somewhat less than 1.0 gain second amplifier (IC12 pins 5, 6 and 7) feeding into the middle of the C-R-C circuit, via R170 (470k).  This second amplifier, with a negative polarity output, driving the middle of the C-R-C network which itself is strung across the inverting amplifier of the first amplifier IC12 pins 2, 3 and 1, means we get positive feedback, but not quite to the gain of 1.0, which would sustain continual oscillation.

When VR6 is CW (clockwise), the gain of the inverting amplifier is around -0.9, because the input resistor is 47k and the feedback resistor is (in AC terms) 47k in parallel with 500k, which is about 0.9 of 47k. The high value (33uF) of C43 in series with the 500k CW value of VR6 would have (according to my rough calculations) a 1/ (2 pi R C) frequency of about 10Hz, so the gain of this amplifier might be a little higher at the BD frequency.

Reducing VR6 by turning it ACW (anti-clockwise) would reduce the gain of this amplifier which reduces (VR6 ACW = 0 ohms) the resonance of the whole circuit, down to zero, or close to zero, since it becomes R164 driving into R169 in parallel with C43.  So this amplifier system's gain can be varied from about -0.90 (VR6 CW) to about -0.000..? (VR6 ACW).

The resonance when VR6 is CW would depend on the value of these resistors:

 R167  R164  VR6  R169  R170

It may also be affected by R161 and C39, going back through R160.

This circuit is rather complex and I wouldn't want to have to write a definitive treatise on exactly what it does.  I doubt the designers did this - I guess they mucked around until they got the sound they wanted, which I think is a perfectly good approach.

Without doing the complete qualitative and quantitative analysis, we can simply tweak the circuit for more gain in the second amplifier, so we get more feedback, which with the phase shifts due to the capacitors, and the inverting nature of the first amplifier, leads to a higher resonance, longer resonant notes and potentially to self-resonance on a continual basis without any trigger from the sequencer.

I wrote about this to Analogue Heaven on 11 November 1996.  I suggested soldering a 390k resistor in parallel with R164 (47k).

This means that the gain of the amplifier with VR6 CW is:

  (47k || 500k)   /  (47k || 390k)

where "||" means "in parallel with".

This is:

   (1 / ( (1 / 47000)  + ( 1 / 500000) )
 / (1 / ( (1 / 47000)  + ( 1 / 390000) )

 ~=   42961  / 41956  ~=  1.024

This should enable continual self-oscillation when VR6 is CW.  If not, try a 330k.

Now, by backing of VR6 from the fully CW position, you can get resonance as long as you like, to the point of self-resonance without any trigger pulse.

The exact amount you would need to turn it to get a given resonance would depend on the values of all these resistors.  With this mod, if you are only interested in the BD not self-resonating, then you can get any decay time you like - thereby eliminating all this stuff about "one TR-808 BD circuit sounding different from another".


Now, to the second question of why the BD gives "erratic" responses as demonstrated in Justin's sound file.

We can imagine the BD circuit being a sine-wave almost resonant oscillator, which is altered in three ways when a trigger pulse arrives.

 1 - The C40, R163, R162 and D53 circuit provides a positive pulse
     to the BD's resonator amplifier's positive input (IC12 pin 1).

     This, in the absence of any other stimulus or current oscillation
     will cause the circuit to oscillate, starting with a positive
     sine-like cycle starting at the time of the trigger pulse.

 2 - Q41 shorts out C38 which means Q42 turns off and remains off until
     R156 charges up C38 enough, after the end of the 1msec trigger
     pulse, to turn Q42 on again.   During this time, Q43 is off, so
     R165 is shorted out, giving a higher resonant frequency.

 3 - There's a very small (0.6 volts, due to D52 turning on) kick to
     the system via C39 and R161 when Q42 turns on, and a big kick in
     the opposite direction when it turns off (which depends on R156,
     C38, the gain of Q42, the temperature of Q42 and the precise
     value of Jupiter's gravitational influence on all these
     components).

Overall, we can think of this as a temporary switch of resonance frequency to a much higher frequency plus two sources of energy which drive the system to ring afterwards - to create a sinusoidal signal of decreasing amplitude, just like the resonance of a simple drum.

If the system is not resonating at all - it has no sinusoidal change going on (that is, it is silent) - then when we get a trigger pulse of a given amplitude (5V normally, but up to about 14V for an accented beat) then we will get a completely predictable ringing waveform thereafter. The amplitude of this ringing waveform will always be the same from one beat to the next - *provided* we allow enough time between the trigger pulses for the resonance to decay to "zero".  In fact, it never decays to absolute zero - but if we let it decay so far that we can't hear it, which means it is, say, 1/1000 the amplitude it was at the start of the note.

*However* if, when the circuit is still resonating, we give it another trigger pulse, then what will happen?

It depends on:

 1 - The amplitude of the trigger pulse.

 2 - The exact part of the cycle the circuit is at in its nearly
     self-resonant decay.

For the sake of simplicity, I will ignore the frequency change which follows the trigger pulse.  We can imagine the trigger pulse being a single, simple, thing which puts "positive" energy into the nearly self-resonant circuit, of let's say "1 volt".  That is: if the trigger pulse arrived when there was no self-oscillation, then the system would ring with a 1V peak (2V peak to peak) sine-wave which then decays at some rate.

Now let's consider two situations when the trigger pulse arrives when the circuit is still resonating from a previous trigger pulse:

 A: The trigger pulse arrives when the circuit is at a +0.3V peak.

 B: The trigger pulse arrives when the circuit is at a -0.3V peak.

Broadly speaking, what will result, is a new "note", with the following
characteristics:

 A: 1.3 volt peak.  This is louder than the first note.

 B: 0.7 volt peak.  This is quieter than the first note.


So the variations in volume, softer and louder, which appear in Justin's recording are due to the trigger pulse arriving at various times in the oscillation.  The trigger pulse is always the same (except of course if accented notes are used, which I guess is the case in this recording) but in the B case, the trigger pulse's energy went into quenching the oscillation, whereas in the A case, it added to the existing energy, and resulted in a louder note than usual.

Listening to Justin's recording, my impression is that any TR-808 would do this if it had the same sequence of trigger pulses and the same self-resonance.  The exact details of which notes were louder and software would depend on the precise timing of the trigger pulses, the precise state of the oscillation cycle at that time - and so the precise values of resistors, the temperature of the components etc.


As I wrote in 1996, the tuning of the BD circuit can be changed by replacing R165 with a 100k pot.  ACW means higher pitches.  CW means lower pitches.  A 100k log pot would be better than a linear one, but either would do the job.  This involves cutting a track on the top side of the board.  There's no easy way of mounting a pot there, but sufficiently small pots can be mounted in the side panels.

Only technicians should work on TR-808s.  There is the danger of static electricity damage to the circuitry, such as zapping the CPU or a memory chip.  Also, there are exposed mains wires inside the machine, so anyone working on the machine while it is powered up risks DEATH.


 - Robin    http://www.firstpr.com.au/rwi/dfish/
           http://www.firstpr.com.au/rwi/tr-808/


Reply | Threaded
Open this post in threaded view
|

Re: TR-808 Bass Drum has variable punch & decay

Robin Whittle
Hi Greg,

You wrote:

> So you are saying that this particular 808 came off the factory line
> with this issue or is this something that happens to random 808's as the
> circuits age?

I am saying that this is the normal behaviour of TR-808s.  It has
nothing to do with age.

The precise decay time of the BD circuit when the decay pot is fully
clockwise will vary from one machine to the next, due to slight
variations in resistor values.  That might change a little over time,
and perhaps with temperature.

However, the behaviour of the trigger pulse's energy sometimes being
boosted by the resonant signal at the time of the trigger pulse and some
times being diminished by this is a natural part of how the circuit
works.  The degree to which this occurs would not be appreciably
affected by alterations in resistor values.

For instance, if the drum circuit has decayed to 50% of its initial
level when the trigger pulse occurs, then (unless the trigger pulse
quenches all oscillation to some degree, which it might) then we would
expect the second note to be anywhere between 50% and 150% of the
strength of the first note.

It would also be possible to use an accented trigger pulse at time A,
generating a signal of, say, 2.0 volts peak at the start of the note .
. . and then, a fraction of a second later, when the signal has decayed
to 1.0 volt peak, to give it another non-accented trigger pulse, which
would ordinarily create a 1.0 volt signal.  Then what happens depends
entirely on which part of the sine cycle the circuit was in when this
trigger pulse arrived.

If it was at its peak, the second sound would begin with a 2.0 volt
signal.  If it was at its trough, then this would exactly cancel the
trigger pulse's energy and there would be no sound after the trigger
pulse.  This is assuming the trigger pulse doesn't quench any of the
oscillation - however, maybe it does.

Theoretically, the same thing could happen with a physical drum - you
could hit the skin at a time when the movement of the skin was in phase
with the hit, or in the opposite phase, or in mid-phase, resulting in
higher, lower or normal sound after the hit.  In practice, the hit of a
drum stick would last for a longer time than the 1msec trigger pulse of
the TR-808 (and TR-606) circuitry.  Also, with a physical kick drum, the
beater would tend to dampen the oscillation of the skin and its resonant
chamber.

So the purity of the electronics gives a much stronger effect than in a
physical drum circuit.

If you want to avoid this, and still have long decays, this could be
done by some kind of modification of the circuit which I have not yet
contemplated.  Probably a better approach would be to use a sampler,
where one note stops the playback of any previous note.

For a given tone and decay setting, and for a given trigger pulse level,
if the circuit is not currently oscillating, the circuit will produce
exactly one sound.  There is no variation due to the white noise
generator (snare snappy sound, tom rumble or the entire handclap sound)
or the 6 square-wave oscillators used for the cymbal and hi-hat
circuits.  So the TR-808 BD is a perfect candidate for sampling.  Just
go through a few favourite settings with different accent levels and you
have the full repertoire of what the circuit does in isolation - with a
trigger pulse which fires up the circuit from zero signal.

  - Robin

Reply | Threaded
Open this post in threaded view
|

Re: TR-808 Bass Drum has variable punch & decay

Justin Morgan
Hey Robin...

Many thanks for such an informative and detailed reply...
Kind of a head rush for me to make complete sense, I'll admit.

So, it sounds like ... if I want a specific behavior for the decay,
I need to specify this to a tech who might be able to make
those adjustments.
I've used a handful of other 808s in the past... where I was able
to fully turn the decay clockwise and get a consistent (for the most part) decay time,
on a "4 on the floor" beat for the 16 steps.
So, circuit wise... I can only believe it's possible for my 808 to achieve this result.

I'll message you off-list about any more details/questions.
Thanks again.

Justin




On Wed, Mar 9, 2011 at 3:21 AM, Robin Whittle <[hidden email]> wrote:
Hi Greg,

You wrote:

> So you are saying that this particular 808 came off the factory line
> with this issue or is this something that happens to random 808's as the
> circuits age?

I am saying that this is the normal behaviour of TR-808s.  It has
nothing to do with age.

The precise decay time of the BD circuit when the decay pot is fully
clockwise will vary from one machine to the next, due to slight
variations in resistor values.  That might change a little over time,
and perhaps with temperature.

However, the behaviour of the trigger pulse's energy sometimes being
boosted by the resonant signal at the time of the trigger pulse and some
times being diminished by this is a natural part of how the circuit
works.  The degree to which this occurs would not be appreciably
affected by alterations in resistor values.

For instance, if the drum circuit has decayed to 50% of its initial
level when the trigger pulse occurs, then (unless the trigger pulse
quenches all oscillation to some degree, which it might) then we would
expect the second note to be anywhere between 50% and 150% of the
strength of the first note.

It would also be possible to use an accented trigger pulse at time A,
generating a signal of, say, 2.0 volts peak at the start of the note .
. . and then, a fraction of a second later, when the signal has decayed
to 1.0 volt peak, to give it another non-accented trigger pulse, which
would ordinarily create a 1.0 volt signal.  Then what happens depends
entirely on which part of the sine cycle the circuit was in when this
trigger pulse arrived.

If it was at its peak, the second sound would begin with a 2.0 volt
signal.  If it was at its trough, then this would exactly cancel the
trigger pulse's energy and there would be no sound after the trigger
pulse.  This is assuming the trigger pulse doesn't quench any of the
oscillation - however, maybe it does.

Theoretically, the same thing could happen with a physical drum - you
could hit the skin at a time when the movement of the skin was in phase
with the hit, or in the opposite phase, or in mid-phase, resulting in
higher, lower or normal sound after the hit.  In practice, the hit of a
drum stick would last for a longer time than the 1msec trigger pulse of
the TR-808 (and TR-606) circuitry.  Also, with a physical kick drum, the
beater would tend to dampen the oscillation of the skin and its resonant
chamber.

So the purity of the electronics gives a much stronger effect than in a
physical drum circuit.

If you want to avoid this, and still have long decays, this could be
done by some kind of modification of the circuit which I have not yet
contemplated.  Probably a better approach would be to use a sampler,
where one note stops the playback of any previous note.

For a given tone and decay setting, and for a given trigger pulse level,
if the circuit is not currently oscillating, the circuit will produce
exactly one sound.  There is no variation due to the white noise
generator (snare snappy sound, tom rumble or the entire handclap sound)
or the 6 square-wave oscillators used for the cymbal and hi-hat
circuits.  So the TR-808 BD is a perfect candidate for sampling.  Just
go through a few favourite settings with different accent levels and you
have the full repertoire of what the circuit does in isolation - with a
trigger pulse which fires up the circuit from zero signal.

 - Robin


Reply | Threaded
Open this post in threaded view
|

RE: TR-808 Bass Drum has variable punch & decay

el macaco
Thanks Robin.
 
Justin, long story short, it soudns like your 808 has a longer decay than the others you have used, so when fully CW it has the cancellation issue.  Just turn down the decay until it stops, it shouldn't take a drastic reduction depending on the tempo and where your kicks are.
 
Also, make sure you are listening to the 808 on a system that can reproduce the bass frequencies, sometimes the kick is still ringing but on small speakers you can't hear the low end of the decay so it sounds like the kick has finished decaying but it has not.  This might also make you think you need to turn the decay control up because you want it to last until the next kick.
 
I actually love this about the 808, and I wouldn't alter the circuit to remove it.  That said, maybe yours is not quite right and you could try replacing the capacitors and or diodes in the BD circuit if you are competant with a soldering iron or know someone who is.
 

 

Date: Wed, 9 Mar 2011 12:07:21 -0500
From: [hidden email]
To: [hidden email]
CC: [hidden email]
Subject: Re: [AH] TR-808 Bass Drum has variable punch & decay

Hey Robin...

Many thanks for such an informative and detailed reply...
Kind of a head rush for me to make complete sense, I'll admit.

So, it sounds like ... if I want a specific behavior for the decay,
I need to specify this to a tech who might be able to make
those adjustments.
I've used a handful of other 808s in the past... where I was able
to fully turn the decay clockwise and get a consistent (for the most part) decay time,
on a "4 on the floor" beat for the 16 steps.
So, circuit wise... I can only believe it's possible for my 808 to achieve this result.

I'll message you off-list about any more details/questions.
Thanks again.

Justin




On Wed, Mar 9, 2011 at 3:21 AM, Robin Whittle <[hidden email]> wrote:
Hi Greg,

You wrote:

> So you are saying that this particular 808 came off the factory line
> with this issue or is this something that happens to random 808's as the
> circuits age?

I am saying that this is the normal behaviour of TR-808s.  It has
nothing to do with age.

The precise decay time of the BD circuit when the decay pot is fully
clockwise will vary from one machine to the next, due to slight
variations in resistor values.  That might change a little over time,
and perhaps with temperature.

However, the behaviour of the trigger pulse's energy sometimes being
boosted by the resonant signal at the time of the trigger pulse and some
times being diminished by this is a natural part of how the circuit
works.  The degree to which this occurs would not be appreciably
affected by alterations in resistor values.

For instance, if the drum circuit has decayed to 50% of its initial
level when the trigger pulse occurs, then (unless the trigger pulse
quenches all oscillation to some degree, which it might) then we would
expect the second note to be anywhere between 50% and 150% of the
strength of the first note.

It would also be possible to use an accented trigger pulse at time A,
generating a signal of, say, 2.0 volts peak at the start of the note .
. . and then, a fraction of a second later, when the signal has decayed
to 1.0 volt peak, to give it another non-accented trigger pulse, which
would ordinarily create a 1.0 volt signal.  Then what happens depends
entirely on which part of the sine cycle the circuit was in when this
trigger pulse arrived.

If it was at its peak, the second sound would begin with a 2.0 volt
signal.  If it was at its trough, then this would exactly cancel the
trigger pulse's energy and there would be no sound after the trigger
pulse.  This is assuming the trigger pulse doesn't quench any of the
oscillation - however, maybe it does.

Theoretically, the same thing could happen with a physical drum - you
could hit the skin at a time when the movement of the skin was in phase
with the hit, or in the opposite phase, or in mid-phase, resulting in
higher, lower or normal sound after the hit.  In practice, the hit of a
drum stick would last for a longer time than the 1msec trigger pulse of
the TR-808 (and TR-606) circuitry.  Also, with a physical kick drum, the
beater would tend to dampen the oscillation of the skin and its resonant
chamber.

So the purity of the electronics gives a much stronger effect than in a
physical drum circuit.

If you want to avoid this, and still have long decays, this could be
done by some kind of modification of the circuit which I have not yet
contemplated.  Probably a better approach would be to use a sampler,
where one note stops the playback of any previous note.

For a given tone and decay setting, and for a given trigger pulse level,
if the circuit is not currently oscillating, the circuit will produce
exactly one sound.  There is no variation due to the white noise
generator (snare snappy sound, tom rumble or the entire handclap sound)
or the 6 square-wave oscillators used for the cymbal and hi-hat
circuits.  So the TR-808 BD is a perfect candidate for sampling.  Just
go through a few favourite settings with different accent levels and you
have the full repertoire of what the circuit does in isolation - with a
trigger pulse which fires up the circuit from zero signal.

 - Robin


Reply | Threaded
Open this post in threaded view
|

Re: TR-808 Bass Drum has variable punch & decay

Robin Whittle
Hi Justin, Ed and All,

I think Ed is right to suggest that Justin's TR-808 might have a longer
decay time than other machines and that if it is monitored in a bass
deficient environment, there may be a temptation to increase the decay
time.

All that is needed is for the speakers or the room to fail to respond
properly to the ~50Hz sine wave (~6.6 metre wavelength ~= 22ft) of the
Bass Drum circuit.   That could easily happen with resonances in a room
and/or with smaller speakers.

I don't think there's any reason to suspect component failure.  I think
Justin's recording sounds like normal 808 behavior with that long decay
time and short times between triggers.

 - Robin
Reply | Threaded
Open this post in threaded view
|

Re: TR-808 Bass Drum has variable punch & decay

dancemachine
In reply to this post by el macaco
I was gonna mention the same thing.  my 808 has a mod to extend the
decay of the kick drum, although I can't say I've really found a use
for it, partially due to this phase thing, and partially cause I just
don't find a need for kick drums nearly as long as this thing will go.

but I could see a creative use for this, by just having the kick drum
extend a tiny bit into the next, giving a nice slight variation with
each hit, and you could augment or diminish the effect with the decay
control.  this is why you can't sample the 808.  ;)



On Wed, Mar 9, 2011 at 12:38 PM, el macaco <[hidden email]> wrote:

> Thanks Robin.
>
> Justin, long story short, it soudns like your 808 has a longer decay than
> the others you have used, so when fully CW it has the cancellation issue.
> Just turn down the decay until it stops, it shouldn't take a drastic
> reduction depending on the tempo and where your kicks are.
>
> Also, make sure you are listening to the 808 on a system that can reproduce
> the bass frequencies, sometimes the kick is still ringing but on small
> speakers you can't hear the low end of the decay so it sounds like the
> kick has finished decaying but it has not.  This might also make you think
> you need to turn the decay control up because you want it to last until the
> next kick.
>
> I actually love this about the 808, and I wouldn't alter the circuit to
> remove it.  That said, maybe yours is not quite right and you could try
> replacing the capacitors and or diodes in the BD circuit if you are
> competant with a soldering iron or know someone who is.
>
>
>
> ________________________________
> Date: Wed, 9 Mar 2011 12:07:21 -0500
> From: [hidden email]
> To: [hidden email]
> CC: [hidden email]
> Subject: Re: [AH] TR-808 Bass Drum has variable punch & decay
>
> Hey Robin...
>
> Many thanks for such an informative and detailed reply...
> Kind of a head rush for me to make complete sense, I'll admit.
>
> So, it sounds like ... if I want a specific behavior for the decay,
> I need to specify this to a tech who might be able to make
> those adjustments.
> I've used a handful of other 808s in the past... where I was able
> to fully turn the decay clockwise and get a consistent (for the most part)
> decay time,
> on a "4 on the floor" beat for the 16 steps.
> So, circuit wise... I can only believe it's possible for my 808 to achieve
> this result.
>
> I'll message you off-list about any more details/questions.
> Thanks again.
>
> Justin
>
>
>
>
> On Wed, Mar 9, 2011 at 3:21 AM, Robin Whittle <[hidden email]> wrote:
>
> Hi Greg,
>
> You wrote:
>
>> So you are saying that this particular 808 came off the factory line
>> with this issue or is this something that happens to random 808's as the
>> circuits age?
>
> I am saying that this is the normal behaviour of TR-808s.  It has
> nothing to do with age.
>
> The precise decay time of the BD circuit when the decay pot is fully
> clockwise will vary from one machine to the next, due to slight
> variations in resistor values.  That might change a little over time,
> and perhaps with temperature.
>
> However, the behaviour of the trigger pulse's energy sometimes being
> boosted by the resonant signal at the time of the trigger pulse and some
> times being diminished by this is a natural part of how the circuit
> works.  The degree to which this occurs would not be appreciably
> affected by alterations in resistor values.
>
> For instance, if the drum circuit has decayed to 50% of its initial
> level when the trigger pulse occurs, then (unless the trigger pulse
> quenches all oscillation to some degree, which it might) then we would
> expect the second note to be anywhere between 50% and 150% of the
> strength of the first note.
>
> It would also be possible to use an accented trigger pulse at time A,
> generating a signal of, say, 2.0 volts peak at the start of the note .
> . . and then, a fraction of a second later, when the signal has decayed
> to 1.0 volt peak, to give it another non-accented trigger pulse, which
> would ordinarily create a 1.0 volt signal.  Then what happens depends
> entirely on which part of the sine cycle the circuit was in when this
> trigger pulse arrived.
>
> If it was at its peak, the second sound would begin with a 2.0 volt
> signal.  If it was at its trough, then this would exactly cancel the
> trigger pulse's energy and there would be no sound after the trigger
> pulse.  This is assuming the trigger pulse doesn't quench any of the
> oscillation - however, maybe it does.
>
> Theoretically, the same thing could happen with a physical drum - you
> could hit the skin at a time when the movement of the skin was in phase
> with the hit, or in the opposite phase, or in mid-phase, resulting in
> higher, lower or normal sound after the hit.  In practice, the hit of a
> drum stick would last for a longer time than the 1msec trigger pulse of
> the TR-808 (and TR-606) circuitry.  Also, with a physical kick drum, the
> beater would tend to dampen the oscillation of the skin and its resonant
> chamber.
>
> So the purity of the electronics gives a much stronger effect than in a
> physical drum circuit.
>
> If you want to avoid this, and still have long decays, this could be
> done by some kind of modification of the circuit which I have not yet
> contemplated.  Probably a better approach would be to use a sampler,
> where one note stops the playback of any previous note.
>
> For a given tone and decay setting, and for a given trigger pulse level,
> if the circuit is not currently oscillating, the circuit will produce
> exactly one sound.  There is no variation due to the white noise
> generator (snare snappy sound, tom rumble or the entire handclap sound)
> or the 6 square-wave oscillators used for the cymbal and hi-hat
> circuits.  So the TR-808 BD is a perfect candidate for sampling.  Just
> go through a few favourite settings with different accent levels and you
> have the full repertoire of what the circuit does in isolation - with a
> trigger pulse which fires up the circuit from zero signal.
>
>  - Robin
>
>
>