[sdiy] emulating bi-polar capacitors

Wed Feb 16 17:29:21 CET 2022

I have no experience with a real 303, but looking at the schematic it is 
easy to see why that would be the case.

C29 smooths the bias voltage rail for the differential pair on the 
output of the VCF. That rail biasses the two diff-amp transistor bases 
but also sets their tail current.

If C29 leaks, the bias rail will drop and so will the tail current. And 
a diff-amp with a variable tail current is actually a very crude sort of 
VCA. So lower bias voltage = lower gain for the diff-amp = lower output 
from the VCF = not so much signal to feed back to generate resonance.

Cheers,

Steve L.
Benden Sound Technology

On 16/02/2022 12:35, rburnett at richieburnett.co.uk wrote:
> Yes, I totally agree.  Out of all the places where they might have put 
> one of the electrolytics in backwards, this is one place where you 
> wouldn't expect them to have got away with it!
>
> I had a bit of a play about at the time when I discovered this... 
> Replacing the cap with a new one fitted either way around had the 
> effect of making the filter sound more slightly more resonant and 
> "shrieky" particularly at the high-frequency end.  But that could just 
> be a new vs old capacitor thing, or even just down to capacitance 
> tolerance.  I never got round to actually measuring C and ESR of the 
> old pulled cap though, before I put the original cap back in (the 
> wrong way around again of course) and closed up the box.  Those old 
> phenolic PCBs don't put up with much soldering/desoldering :-(
>
> My impression is that leakage current through the cap fitted the wrong 
> way round was decreasing the loop gain around the filter and making it 
> less resonant.  So it might account for the fact that the TB-303 
> filter never quite manages self-oscillation without some other 
> modification.  But as you said, it would be a strange way to guarantee 
> this across a production run and for the lifetime of the product!
>
> -Richie,
>
>
>
>
>
> On 2022-02-16 11:57, Steve Lenham wrote:
>> That's less an endorsement of playing fast and loose with
>> electrolytics than a case of marvelling at how they got away with it
>> at all!
>>
>> C29 in conjunction with R116 (10K) is acting as a lowpass filter,
>> decoupling the 5V bias voltage being fed to the VCF.
>>
>> If R116 was much lower in value - or not there at all - how long do we
>> think C29 would last?
>>
>> The unpredictable reverse leakage current through C29 probably means
>> that the bias voltage varies quite a bit from unit to unit and perhaps
>> with other things like temperature too. By all means claim
>> retrospectively that this contibutes to the magic of the TB303 but I
>> would be amazed if it were deliberate and it isn't good engineering.
>>
>> Ah, you might say - good engineering doesn't matter if it sounds
>> right. But what we don't know is how many units got binned because the
>> leakage was so bad that the VCF didn't work at all...
>>
>> Cheers,
>>
>> Steve L.
>> Benden Sound Technology
>>
>>
>> On 16/02/2022 10:42, rburnett at richieburnett.co.uk wrote:
>>> There is at least one polar electrolytic capacitor in the Roland 
>>> TB-303 that is fitted backwards and has significant average voltage 
>>> across it (the wrong way!)...
>>>
>>> https://www.matrixsynth.com/2010/11/component-c29-is-backwards-on-all-tb.html 
>>> The polarity of the capacitor is correct on the schematic in the 
>>> service manual, but the PCB ident is the wrong way round. The actual 
>>> component is fitted as per the PCB ident, not the schematic!
>>>
>>> -Richie,
>>>
>>>
>>>
>>> On 2022-02-16 10:00, Steve Lenham wrote:
>>>> just because we can "get away" with using standard polar electrolytics
>>>> (and we often can) doesn't mean that it is good engineering or that
>>>> there will be no negative effects under any circumstances.
>>>>
>>>> Polar electrolytics will withstand something like half a volt of
>>>> reverse polarity with no ill effect. Above that, they start to leak
>>>> current, and if that current is not controlled then they will suffer
>>>> damage up to and including exploding. We've all seen what happens to
>>>> PSU caps accidentally fitted backwards.
>>>>
>>>> In the case of audio coupling caps, the cap forms a single-pole
>>>> highpass filter with the input resistance of the stage it's feeding,
>>>> i.e. a reactive/resistive potential divider. At frequencies well above
>>>> the corner frequency of the filter, virtually all of the voltage
>>>> appears across the resistance. Great, no voltage across the cap (in
>>>> either direction!). But what happens as the frequency drops towards
>>>> the corner frequency? More and more of the voltage appears across the
>>>> capacitor until eventually the negative signal peaks cause the cap to
>>>> leak current. At that point you will start to get signal distortion.
>>>> The cap will rarely be damaged because there is usually enough
>>>> resistance in the circuit to limit the current to a safe level, but
>>>> fidelity might be.
>>>>
>>>> That is why, as previously mentioned, you see bipolar electrolytics in
>>>> speaker crossovers. The corner frequencies of the filters in a
>>>> crossover, by definition, are located in the middle of the audio band,
>>>> so there will be substantial signal voltage across the capacitors at
>>>> all times. In addition, a speaker is usually a low-impedance system,
>>>> so there is nothing to limit any reverse leakage currents to a safe
>>>> level. Polar electrolytics would be a disaster there.
>>>>
>>>> So in a nutshell, polar electrolytics for coupling are fine as long as
>>>> you can be sure that there will never be any substantial signal
>>>> content below the corner frequency of the filter that the cap forms
>>>> with the input impedance of the stage it's feeding.
>>>>
>>>> Of course, this raises the question of what the cap is doing there at
>>>> all. If the above condition is met, the cap isn't doing any highpass
>>>> filtering. And it can't block unpredictable DC offsets (i.e. ones that
>>>> could be of either polarity) greater than half a volt. The answer is
>>>> that it is generally blocking small opamp offset voltages (of the
>>>> order of tens of millivolts) and preventing their being amplified by
>>>> subsequent gain stages.
>>>>
>>>> Cheers,
>>>>
>>>> Steve L.
>>>> Benden Sound Technology
>>>>
>>>>
>>>> On 16/02/2022 06:30, Brian Willoughby wrote:
>>>>> While writing firmware for a digital mixer, the folks over in the 
>>>>> digital-controlled-analog-preamp team discovered distortion that 
>>>>> was traced to the polarized capacitors. I did not look at the 
>>>>> schematic, so I don't know what the exact problem was. They 
>>>>> changed the capacitors for the next prototype and got rid of the 
>>>>> distortion. I don't remember whether they switched to bi-polar 
>>>>> caps or just found polar caps with a higher voltage rating. I seem 
>>>>> to recall it was the latter, but it stuck in my mind that the 
>>>>> electrolytics could cause distortion due to their polarized nature 
>>>>> - at least for large input signals and/or high gain settings.
>>>>>
>>>>> Sorry for the non-answer. Seems like there's definitely the 
>>>>> potential for problems that can't be ignored, but the solution is 
>>>>> not necessarily bi-polar.
>>>>>
>>>>> Brian
>>>>>
>>>>>
>>>>> On Feb 15, 2022, at 12:30, Mattias Rickardsson wrote:
>>>>>> Hej Danjel and others,
>>>>>>
>>>>>> I don't recall seeing any hard facts about non-polarized 
>>>>>> electrolytics performing better than ordinary polarized in audio 
>>>>>> circuits, but still they occasionally turn up in designs. Would 
>>>>>> be interesting to hear why they sometimes are preferred by audio 
>>>>>> designers, though! :-)
>>>>>>
>>>>>> I searched for "non-polarized" in Douglas Self's reference book 
>>>>>> "Small Signal Audio Design" and found a couple of applications 
>>>>>> where they actually do make sense - but it's a practical reason 
>>>>>> rather than an audio performance reason:
>>>>>> DC blocking in inputs & outputs, where it's possible that the 
>>>>>> connected gear pulls the voltage way off ground level, and you 
>>>>>> never know in what direction.
>>>>>> Douglas Self writes:
>>>>>>
>>>>>> "C2 is a DC-blocking capacitor to prevent voltages from 
>>>>>> ill-conceived source equipment getting into the circuitry. It is 
>>>>>> a non-polarized type as voltages from the outside world are of 
>>>>>> unpredictable polarity, and it is rated at not less than 35 V so 
>>>>>> that even if it gets connected to defective direct-coupled 
>>>>>> equipment with an op-amp output jammed hard against one of the 
>>>>>> supply rails, no harm will result."
>>>>>>
>>>>>> Any other good reasons for using them?
>>>>>> And sorry, no - I don't have any better advice than what has 
>>>>>> already been said. :-)
>>>>>>
>>>>>> /mr
>>>>>>
>>>>>>> On Feb 15, 2022, at 11:32 AM, Mike Bryant wrote:
>>>>>>>> I've never bothered, I've never noticed a difference in just 
>>>>>>>> using a 47μ instead of playing about with back-to-back 
>>>>>>>> capacitors even after the thick end of 40 years the earliest 
>>>>>>>> stuff I built that way still has capacitors that capacitate 
>>>>>>>> just fine.
>>>>>>> -- 
>>>>>>>> Gordonjcp
>>>>>>> Agreed.  Unless you actually have a reverse DC bias voltage (in 
>>>>>>> which case rotate the capacitor) I've never understood any need 
>>>>>>> for the non-polarised capacitors. Most mixing consoles are full 
>>>>>>> of thousands of them either feeding the input or fed from the 
>>>>>>> output of an opamp via a resistor to ground.
>>>>>>>
>>>>
>>>> _______________________________________________
>>>> Synth-diy mailing list
>>>> Synth-diy at synth-diy.org
>>>> http://synth-diy.org/mailman/listinfo/synth-diy
>>>> Selling or trading? Use marketplace at synth-diy.org
>>> _______________________________________________
>>> Synth-diy mailing list
>>> Synth-diy at synth-diy.org
>>> http://synth-diy.org/mailman/listinfo/synth-diy
>>> Selling or trading? Use marketplace at synth-diy.org
>>
>>
>> _______________________________________________
>> Synth-diy mailing list
>> Synth-diy at synth-diy.org
>> http://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace at synth-diy.org