Okay, working away on this (more to come once I get it done) and I had a thought: he has a 1k resistor on the led to prevent it from burning out, and there’s a 1k resistor on the tip of the audio jack, which I’m guessing is either buffering the output or maybe they’re pulldown resistors (along with the diodes). Either way, they’re wired together then to the pins through the diodes.
So, my thought was, couldn’t I just use a 2k resistor from the LED cathode to the tip of the output and then wire from the tip to the board connection? Is there a compelling reason not to do this?
Yes.
The resistors are current limiters. The one on the LED takes the output voltage and converts it to a finite current which governs the brightness of the LED. (Use a higher resistance if you want the LED less bright.) The one on the jack prevents shorting the ADC output if the jack gets connected to ground or a voltage rail, which would be bad for the ADC. Even with the resistor it’d still be somewhat bad, probably, if the output were shorted to the negative rail, that’s one of the things I’d want to add protection for.
As I interpret your question you’re asking about something like this:
I assume you meant anode where you said cathode, and I added a 1k between the ADC and jack for the above mentioned reason. But then the two resistors form a voltage divider that reduces the voltage level at the output. (Also the resistance to the LED is now 3k, so it’ll be significantly dimmer.)
Or if you meant something like this:
that would be even worse. The voltage at the jack would be the same as the voltage at the LED anode, which when the ADC output is high would be approximately only 700 mV above ground — the rest of the ADC output voltage having been dropped across the resistor.
As originally drawn, the two resistors keep the LED anode voltage and the jack voltage separate so the one doesn’t adversely affect the other.
3 Likes
Ah, thanks @analogoutput ! That was really helpful (as always 
).
Here’s the end result:
I measured out the holes for the jacks and LEDS initially, then, after the drill bit slipped three times in a row, I just decided to lean into the wonky layout.
It works well based on my tests but the perfboard I used isn’t super great quality so I don’t know how long it will last. Overall the process was mostly painless and it’s a fun way to add some variety to patches.
2 Likes
Okay, so, revisiting this and reading up on OP Amp buffering, I came across this article, which - in addition to the standard op amp buffer - suggests a transistor buffer.
I’m curious to get feedback on this as a buffer for this module. The author of the article has some thoughts:
That’s a very good question! The transistor version can be made very compact and squeezed into odd gaps on a busy board if needed. The part types and component values are not critical in this transistor circuit, and it will give a good voltage level output for any positive gate input level. Getting a supply-level output in the op-amp version requires selecting parts correctly. On the other hand, the op-amp circuit here can be tweaked to have a higher threshold, and the simple comparator can be easily changed into a Schmitt trigger, which might prove useful with unstable gate inputs. It might also provide a better source for multiple parallel triggers. For most purposes, I think it comes down to personal choice, and most of the time these days I’d prefer the transistor circuit unless I had an op-amp stage spare somewhere already.
Any thoughts, suggestions, opinions, recons, etc are eagerly anticipated.
I am planning to build a 123ADC module to trigger some drum modules. I see that the stripboard layout uses a 7805 voltage regulator to provide 5V to power the ADC0820. I assume this also means the trigger output voltage is 5V. Is the 5V enough to trigger most drum modules?
The datasheet of the TI ADC0820-N ( ADC0820 8-Bit High Speed uP Compatible A/D Converter with Track/Hold Function datasheet (Rev. C)) mentions under the ‘Absolute Maximum Ratings’ a max voltage of 10V, and under ‘Operating Ratings’ it mentions 4,5V to 8V.
Because of all this, would it be worthwhile to power the ADC with (for instance) 6V of 7V (using a 7806/7807), in order to have a higher trigger output voltage, thereby increasing the possibility that the module will trigger more drum modules?
Thanks for everyone’s input!
I think most drum modules will trigger on 5 V. I know the Barton ones trigger on quite low voltages, a few hundred millivolts I think; in fact in my experience the problem is they sometimes are too trigger happy.
I’ve thought about building something based on the 123ADC, and one change I’d make is to add buffering to the outputs to protect the ADC. Transistors or op amp voltage followers. These could also optionally do logic to AND the outputs with the clock if you want pulsed outputs. If the buffers are powered from ±12 V they can step the outputs up to wherever you need them to be. 5 volt regulators are easier to come by than 6 or 7 volt and the ADC is designed for best performance with a 5 V supply.
2 Likes
Good to hear that most drum modules seem to trigger at 5V. I have made a Kick and Snare modules based of SoundForce PCBs; I will do some tests to make sure they are triggered by 5V or less. If so, then there is no need to mess around with the trigger voltage of the 123ADC.
I like your suggestion of protecting the ADC IC! I think I have some additional 2N3904 or BC547 NPN transistors to use for this.
1 Like
This got me thinking: if you have the ADC outputs going into a DAC, does this make some sort of Sample & Hold module? If so, do the clock rates going into the ADC and DAC have to be the same, or are different clock rates possible, leading to some ‘funky’ irratic behaviour?
Yeah, the ADC output stays fixed until a clock pulse comes in to tell it to do a new digitization, and the DAC output stays fixed until a clock pulse comes in to tell it to do a new conversion, so it can work as a sample and hold (though the output is quantized to 256 values — or for instance 128 if you suppress the lowest bit… which could be useful as a simple pitch quantizer). The clock rates could be different, but if the DAC clock is faster it just gets the same conversion until the ADC clock causes the ADC output to change, and if it DAC clock is slower it potentially skips over one or more ADC outputs — either way I don’t see offhand much benefit.
I’ve thought about building a module with both an ADC and a DAC, there are some potentially fun things you could do, like scrambling the order of the ADC output bits into the DAC inputs to make a “sample and hold” whose input maps to its output in some complicated way.
1 Like
Thanks!
Would it be possible to use the ADC+DAC also as a CV quantizer? If this is possible, would it be stable enough to work as a 1v/Oct CV quantizer? If so, why am I not able to find a schematic for this? 
In principle it should work. Assuming 8 bit ADC and DAC, if you design it so 10 volts at the input are scaled to give you 240 ADC counts and 240 into the DAC gives you 10 volts at the output, and then suppress the least significant bit, then it’ll quantize to 120 levels in 0 to 10 volts, which at 1 V/oct corresponds to semitones. You’d have to be careful with the scaling to get it accurate. I think most quantizers use something like an Arduino and a DAC, and of course the Arduino has an ADC so that’s kind of exactly how it works, except that the microcontroller handles the communications between the ADC and the DAC and can do things like impose major/minor scales etc. There’s also the question of how linear the ADC and DAC are, with 8 bits they might not be as good as you’d like; Arduino ADC is 10 bits and you often see 12 bit DACs in these things.
1 Like
Before making a S&H, I want to build a 123ADC. I am planning to make one in a 10x10cm panel, which only has room for 6 of the 8 outputs (+LEDs).
Two questions:
Would it be better to lose the first two outputs than the last outputs? Otherwise the level LED would not be linked to all outputs high.
Should the unused outputs of the ADC be tied to ground?
Not unambiguous what you mean by “first” and “last”, but if I were constrained to six outputs I’d use the six high order bits, i.e. pins 4, 5, 14, 15, 16, 17. What this amounts to is using this eight bit ADC as a six bit ADC. The range from output 000000 to 111111 will still correspond to voltage range 0 V to Vref, just at reduced resolution. If you use the low order bits the range from output 000000 to 111111 will correspond to voltage range 0 V to Vref/4, probably not what you want, unless you make other modifications.
In Sam’s layout, the one LED that isn’t associated with a single output is an overflow indicator, meaning the input level is higher than the reference voltage. If you use the six low order bits, that LED won’t light until the input voltage is more than four times the voltage needed to produce 111111.
I wouldn’t do that unless the datasheet said to, and it doesn’t (nor does it say not to).
2 Likes
Thanks Richard! Appreciate all your input!
