ARP 1601 clone Build Thread

[fprs87]

the quantizer is giving me some trouble. i have the sequencer stopped and step 1 active. the quantizer A output range is 0-2V, but bounces around quite a bit when i move the slider. it looks like it dips to 0V and then comes back to where the slider is set. it also changes notes without slider movement irregularly every 20 clock cycles or so. the note change seems to happen only when the slider is set <50%. the unquantized output goes from 0-10V.

I'm having the exact same issue, but it only happens when the sequencer is set to 16/1 and both Quantized outputs jump to 0V in between every chromatic half step. Works fine in 8 step.

Did you end up solving this issue? I ended up replacing Z21, Z24, Z25, and Z27 as Kipling suggested along with a continuity test and it seems to check out, but the problem still persists.

I'm having these issues too. v6a board 1.6.2 BOM

TP4 sawtooth is the same as yours, does not go all the way down to 0v before resetting.
16/1 mode the quantized output jumps down to 0v inbetween steps. In 8/2 mode the first 50% of some sliders nothing happens then will quantize after 50%, though this seems to be a bit intermittent. ICs Z21,Z24,Z25,Z27 are checked in a little IC tester and are good.

The rest of the test points seem to look ok as in the service manual.

[CollinRussell]

I'm having these issues too. v6a board 1.6.2 BOM

TP4 sawtooth is the same as yours, does not go all the way down to 0v before resetting.
16/1 mode the quantized output jumps down to 0v inbetween steps. In 8/2 mode the first 50% of some sliders nothing happens then will quantize after 50%, though this seems to be a bit intermittent. ICs Z21,Z24,Z25,Z27 are checked in a little IC tester and are good.

The rest of the test points seem to look ok as in the service manual.

Verbatim - the exact same issues I'm having.
Please keep us all updated if you find the culprit and I'll do the same.

[elmegil]

The TP4 thing is "normal". I have a note in my copy of the service manual that it has about 7 / 7.5V offset from what is shown and the one I have in front of me shows the same behavior but works 100%. Kind of wonder if It's not a mistake in the manual. The thing about the quantizer though, that's weird. The one I have complete on my bench right now works fine in both modes (16x1, 8x2) with the sliders low, and I can see all the voltages varying.

.... I wonder if the fact that it's "some sliders" and intermittent might mean there's a solder joint issue?

Did you build from the synthcube parts kit or did you get the BOM yourself? If yourself, what are the exact part numbers and manufacturers for those chips? I don't *know* this to be a problem, but some CMOS chips from different manufacturers behave differently, and we can compare notes. Of course if it's the Synthcube kit, which is what I'm building from, odds are very low that mine is different from yours, but it might still be worth checking.

I have a working model here in front of me, so if you want to take measurements at different places besides just the test points I can tell you what they look like on my working one.

[elmegil]

One other thing -- have you already calibrated, or are you trying to troubleshoot first? I've found with these there are times when stuff isn't working, but starts doing so once it's calibrated. Most commonly I've seen that with the clock speed (before calibration, lowest clock speed doesn't run the oscillator at all). I don't think that's necessarily any part of the quantizer problem, but it's a good thing to understand we we're going forward.

[CollinRussell]

The TP4 thing is "normal". I have a note in my copy of the service manual that it has about 7 / 7.5V offset from what is shown and the one I have in front of me shows the same behavior but works 100%. Kind of wonder if It's not a mistake in the manual. The thing about the quantizer though, that's weird. The one I have complete on my bench right now works fine in both modes (16x1, 8x2) with the sliders low, and I can see all the voltages varying.

.... I wonder if the fact that it's "some sliders" and intermittent might mean there's a solder joint issue?

Did you build from the synthcube parts kit or did you get the BOM yourself? If yourself, what are the exact part numbers and manufacturers for those chips? I don't *know* this to be a problem, but some CMOS chips from different manufacturers behave differently, and we can compare notes. Of course if it's the Synthcube kit, which is what I'm building from, odds are very low that mine is different from yours, but it might still be worth checking.

I have a working model here in front of me, so if you want to take measurements at different places besides just the test points I can tell you what they look like on my working one.

I got my kit from Synth Cube - PCB revision V6a - and all IC's are matched with part numbers on the Synthcube BOM 1.6.1.
Good to see I'm not crazy about TP4 - So you verified that the sawtooth wave is clipped at the top and offset by around +7V

Solder joints look good on all faders and the mode switch (I have a lot of experience soldering these faders and soldered, OA'd and QC'd professionally for about 6 years).
I've replaced Z21, Z24, Z25, and Z27 and checked for continutity as per Kipling's suggestion.
Reflowed anything that looks even remotely suspicious.
The unit is fully calibrated.

To sum up:

1. In 16/1 mode, every slider drops to 0V between every chromatic half step. The slider's can't decide what voltage to be set to, so when the sequencer loops around, sometimes its the value I set and other times it's 0V (as if the slider is fully down).
2. Every slider is intermittent up until around 50% when in 8/2 mode, but there is not a 0V drop in between chromatic half steps. Sometimes this mode works perfectly in terms of pitch.
3. I can only calibrate the maximum clock rate to 86Hz (the service manual specifies 100Hz) with the multi turn trimmer provided with the Rev 6A BOM. That's the end of of the trim pot's range.
4. When calibrating the Mod Adjust for Quantizer A and B, the sequencer output peaks at +4.235V when the service manual says it "should be near +5V". Is that near enough?
5. When in 8/2 mode and neither Reset or Skip is set, the Gate Bus switches of either 8 step sequencer control the gate patterns of both 8 step sequencers. I am unable to have different gate patterns per 8 step sequencer. Couldn't find anything about this bevahior in the manual or service manual so I'm unsure if this is normal or not. Seemed curious.


S.O.S.!

Thanks for reply!

[fprs87]

The TP4 thing is "normal". I have a note in my copy of the service manual that it has about 7 / 7.5V offset from what is shown and the one I have in front of me shows the same behavior but works 100%. Kind of wonder if It's not a mistake in the manual. The thing about the quantizer though, that's weird. The one I have complete on my bench right now works fine in both modes (16x1, 8x2) with the sliders low, and I can see all the voltages varying.

.... I wonder if the fact that it's "some sliders" and intermittent might mean there's a solder joint issue?

Did you build from the synthcube parts kit or did you get the BOM yourself? If yourself, what are the exact part numbers and manufacturers for those chips? I don't *know* this to be a problem, but some CMOS chips from different manufacturers behave differently, and we can compare notes. Of course if it's the Synthcube kit, which is what I'm building from, odds are very low that mine is different from yours, but it might still be worth checking.

I have a working model here in front of me, so if you want to take measurements at different places besides just the test points I can tell you what they look like on my working one.

Thanks for the reply!

I'm building from the synthcube parts kit, component packages are all labeled with V6a 1.6.2. I read that matching Q4/5 wasn't needed so I didn't try.
I was able to calibrate, though in A MOD ADJUST / B MOD ADJUST the sequencer output was around 4.528V, where the note mentions it should be near +5V, close enough I hope?

The unquantized outputs of each slider goes smoothly from 0-10V which tells me the sliders themselves are good, but from the quantized outputs is where it gets spotty, with only the upper 50% of the travel range working in 8/2 mode or dropping to 0v between notes in 16/1. Something wrong going on in the quantizer circuit I believe.

TP-18 I can see the stairstepped waveform from 0-2V so I guess that means the resistor chain is ok. What is the output of Z30 pins 9 and 10 supposed to look like? From what I understand in the schematic Z30 supplies the quantized control voltage to the Channel A/B Memory, so should these be a quantized voltage equal to the currently actively slider? Z30 (CD4016BE) pins 9 and 10 look like sawtooths on mine

[elmegil]

I got my kit from Synth Cube - PCB revision V6a - and all IC's are matched with part numbers on the Synthcube BOM 1.6.1.
Good to see I'm not crazy about TP4 - So you verified that the sawtooth wave is clipped at the top and offset by around +7V

yes indeed it is....

Solder joints look good on all faders and the mode switch (I have a lot of experience soldering these faders and soldered, OA'd and QC'd professionally for about 6 years).
I've replaced Z21, Z24, Z25, and Z27 and checked for continutity as per Kipling's suggestion.
Reflowed anything that looks even remotely suspicious.
The unit is fully calibrated.

To sum up:

1. In 16/1 mode, every slider drops to 0V between every chromatic half step. The slider's can't decide what voltage to be set to, so when the sequencer loops around, sometimes its the value I set and other times it's 0V (as if the slider is fully down).

So ... I apologize I'm still not quite getting it.

So a half step is 1/12V or roughly 83mV. You're saying as you bring the slider up, instead of going 0V, 83mV, 166mV, 249mV etc, it's 0V, 83mV, 0V, 249mV... or my zeroes are off by one ? Or are you saying that when it's not RIGHT at the exact chromatic half step the quantizer outputs zero, but when you get to the right point it actually outputs the right voltage?

2. Every slider is intermittent up until around 50% when in 8/2 mode, but there is not a 0V drop in between chromatic half steps. Sometimes this mode works perfectly in terms of pitch.

When you say intermittent... you mean sometimes it's the value, and sometimes it's just zero?

What is your power supply (wall wart)? Since you bought the Synthcube kit I assume it's the one that came with it? The Triad WSU150-0800?

3. I can only calibrate the maximum clock rate to 86Hz (the service manual specifies 100Hz) with the multi turn trimmer provided with the Rev 6A BOM. That's the end of of the trim pot's range.

This sounds familiar to me, but I can't remember off hand when I saw it (this one that I have built I finished over a month ago...). I will double check some resistor values there and get back to you.

4. When calibrating the Mod Adjust for Quantizer A and B, the sequencer output peaks at +4.235V when the service manual says it "should be near +5V". Is that near enough?

I will measure that for you tomorrow. I definitely remember not making it to 5V, but I thought it was a bit closer.

5. When in 8/2 mode and neither Reset or Skip is set, the Gate Bus switches of either 8 step sequencer control the gate patterns of both 8 step sequencers. I am unable to have different gate patterns per 8 step sequencer. Couldn't find anything about this bevahior in the manual or service manual so I'm unsure if this is normal or not. Seemed curious.

I will check that as well, I'm not certain. I do think that in 8/2 mode that a switch in, say gate bus 3 for either side on the active step will cause a pulse on the GB3 output. So like a logical OR of the two sets of switch settings. But that's off the top of my head, let me verify.

[elmegil]

One other question: how does the unquantized output look? Aside from being higher voltages, of course.

[CollinRussell]

So ... I apologize I'm still not quite getting it.

So a half step is 1/12V or roughly 83mV. You're saying as you bring the slider up, instead of going 0V, 83mV, 166mV, 249mV etc, it's 0V, 83mV, 0V, 249mV... or my zeroes are off by one ? Or are you saying that when it's not RIGHT at the exact chromatic half step the quantizer outputs zero, but when you get to the right point it actually outputs the right voltage?

That is correct. 0V, 83mV, 0V, 166mV, 0V, 249mV... etc, etc - But only in 16/1 mode. Even when it is on the exact chromatic step, it drifts away from that step every once in a while as the sequencer loops. So if I have the slider set at 166mV, sometimes it outputs 166mV and other times it outputs 0V. I haven't been able to find a sweet spot where it's always exact with every sequencer cycle. So its impossible to know if a step is set correctly until after it fires.

When you say intermittent... you mean sometimes it's the value, and sometimes it's just zero?

Yes, sometimes in 8/2 mode, the entire bottom half of the slider's travel is 0V. Once it reaches its half way point, it jumps to a higher voltage (like 1V) and starts moving chromatically from there.

What is your power supply (wall wart)? Since you bought the Synthcube kit I assume it's the one that came with it? The Triad WSU150-0800?

Correct

This sounds familiar to me, but I can't remember off hand when I saw it (this one that I have built I finished over a month ago...). I will double check some resistor values there and get back to you.

That would be great! Thank you so much. Not the biggest of the issues I'm dealing with and probably within tolerance.

I will measure that for you tomorrow. I definitely remember not making it to 5V, but I thought it was a bit closer.

Again, thank you!

I will check that as well, I'm not certain. I do think that in 8/2 mode that a switch in, say gate bus 3 for either side on the active step will cause a pulse on the GB3 output. So like a logical OR of the two sets of switch settings. But that's off the top of my head, let me verify.

That's what it seems like is happening. Just doesn't seem to be documented anywhere.
I can check the quantized output tomorrow, as I'm forgetting which test point it was in the service manual.
Sad to see that another wiggler is experiencing the exact same issues as I am, but happy that we're in it together!
Hopefully this gets resolved soon. I was lucky enough to use an original 1601 with two 2600s during my time at Berklee, and have been wanting the 1601 in my system ever since - especially after completing my TTSH early last year. Excited for it to be done!

[fprs87]

I made a video to show the faults I'm having, I thought that the 8/2 problem was intermittent but found the setting of steps 9-16 affect the range of steps 1-8 (and vice versa when using the B Quantized output)

Also shows 16/1 mode where steps jump between 0 and quantized note, even though the unquantized outputs work correctly.



@CollinRussell this is pretty similar to what you are experiencing right?

[KSS]

Jumping to zero means the quantizer is beng reset. Differences between a and b means the channel selector is not working correctly. That both of these meet in the middle of the original service manual schematic and share some common glue logic should lead to clues for what needs to be done.

What 'decides' to reset the quantizer clock -and therefore the ramp of notes drawn from for output?
What changes the quantizer between its a and b channels?

[CollinRussell]

I made a video to show the faults I'm having, I thought that the 8/2 problem was intermittent but found the setting of steps 9-16 affect the range of steps 1-8 (and vice versa when using the B Quantized output)

Also shows 16/1 mode where steps jump between 0 and quantized note, even though the unquantized outputs work correctly.



@CollinRussell this is pretty similar to what you are experiencing right?

This is exactly what I’m experiencing. 100% all the way through. I just didn’t realize that the second sequencer slider positions is what was messing up the range of the first sequencer when in 8/2 mode. Good catch.

[CollinRussell]

Jumping to zero means the quantizer is beng reset. Differences between a and b means the channel selector is not working correctly. That both of these meet in the middle of the original service manual schematic and share some common glue logic should lead to clues for what needs to be done.

What 'decides' to reset the quantizer clock -and therefore the ramp of notes drawn from for output?
What changes the quantizer between its a and b channels?

Thanks for always allowing the opportunity to learn. I’ve seen your posts through several threads and am very thankful for your time.

So something is resetting the quantizer counter which is Z25 (CD4024) and something is messing with the channel switch which is Z30 (CD4016). Is that correct?

Still wrapping my head around the circuit and trying to learn this level of debugging at the same time.

Thanks for your patience!

[elmegil]

I apologize, it's been a long day, I haven't forgotten you but I have no steam left. I will try to do the promised measurements tomorrow.

[fprs87]

Jumping to zero means the quantizer is beng reset. Differences between a and b means the channel selector is not working correctly. That both of these meet in the middle of the original service manual schematic and share some common glue logic should lead to clues for what needs to be done.

What 'decides' to reset the quantizer clock -and therefore the ramp of notes drawn from for output?
What changes the quantizer between its a and b channels?

Thanks for the tip, I took a closer look under the scope and in the 16/1 and when it jumps to zero there is an extra reset and sample pulse. This is TP-18 scanner ouput and TP-21 reset under the scope:



I also saw the comparator looks different to how it should in the service manual. Here is TP-18 scanner output and TP-19 comparator. My comparator changes shape as I move the slider and is 350usec duration, where manual says it should be 60-130usec for A/B channels

[fprs87]

I also saw the comparator looks different to how it should in the service manual. Here is TP-18 scanner output and TP-19 comparator. My comparator changes shape as I move the slider and is 350usec duration, where manual says it should be 60-130usec for A/B channels

UPDATE: Since the comparator wave shape looked a bit off, I took out Z23 (LM301) to test it, and it appeared working in an opamp tester I got from Synthchaser.

I tried a different LM301 anyway, pulled from an old OB-8 board I have, and it works! The 16/1 mode no longer jumps to 0V and 8/2 is able to sequence 2 independent pitches.

@CollinRussell if you have another LM301 around you could try swapping it out and see if it solves your issue too?

[synthcube]

Watching all of this with fascination.
We have had some trouble with a small number of LM301AN (Rochester parts) on another project (Pharmasonic Sys-100 VCO) , so we ordered some LM301AP, and that resolved the issue. If you do not have an extra LM301 around, contact us and we will send you LM301AP to try out.

[CollinRussell]

UPDATE: Since the comparator wave shape looked a bit off, I took out Z23 (LM301) to test it, and it appeared working in an opamp tester I got from Synthchaser.

I tried a different LM301 anyway, pulled from an old OB-8 board I have, and it works! The 16/1 mode no longer jumps to 0V and 8/2 is able to sequence 2 independent pitches.

@CollinRussell if you have another LM301 around you could try swapping it out and see if it solves your issue too?

I’ll acquire one and try it!! Thanks keep us posted if anything else comes up

[CollinRussell]

Watching all of this with fascination.
We have had some trouble with a small number of LM301AN (Rochester parts) on another project (Pharmasonic Sys-100 VCO) , so we ordered some LM301AP, and that resolved the issue. If you do not have an extra LM301 around, contact us and we will send you LM301AP to try out.

I’ll need one. I’ll email Gabriel at your shop and set it up. Thanks so much!

[KSS]

UPDATE: Since the comparator wave shape looked a bit off, I took out Z23 (LM301) to test it, and it appeared working in an opamp tester* I got from Synthchaser.

Good news!

Of note in the original schematic is the 10% 3p3 compensation cap on the 301. That's an unusually tight spec compared <--Pun intended! to other 301's in typical ARP circuits. Considering that it is also a compensation capacitor, we might want to play with the caps values in that position before consigning the problem to a certain 301 variant. Maybe the new layout changed the inherent circuit characteristics around this componwnt and 3p3 is no longer the ideal compensation? 3p3 is already a small value. 2p97 to 3p63 being the original expected range.

Dealing with these types of component 'edge cases' are part of making a robust clone.

*That it worked in the OPA tester is another clue it may be more about a compensation 'edge' value than an actual 'bad' component issue.
--------------------
@Synthcube I suspect this may also be true for the 700 clone issues.

[Kipling]

Thanks @KSS. It’s odd that this issue has not occurred until now. There have been no changes to the track layout in that area since my first prototypes which continue to work fine.

It would therefore suggest that component specs/quality may have changed along the way - the original Panasonic MLCC caps became obsolete so had to be substituted with their replacement range but I don’t know if that is the cause here.

I have prepared a Rev 7 layout which changes the obsolete THT LM301AN for the readily available SMT LM201, plus reducing the board from a 4-layer (two of which are power planes) to a 2-layer with routed +15V and copper-pour ground plane, which will make the board cheaper. I haven’t got as far as ordering a minimum order sample for a test build though, partly because I already have my two original prototypes and I wouldn’t cover my costs in doing so unless I were able to sell it on (without front panel or case) assuming it all works as hoped.

[elmegil]

So I'm going to take it that since it appears that the LM301 is at fault, I'm not going to do those measurements :-D

[Kipling]

So I'm going to take it that since it appears that the LM301 is at fault, I'm not going to do those measurements :-D

Looks that way for now. I shall be experimenting with an SMT LM201 shortly, along with a few other changes for a forthcoming Rev 7. They way things are going, it’ll end up being entirely SMT eventually!

[elmegil]

So I'm going to take it that since it appears that the LM301 is at fault, I'm not going to do those measurements :-D

Looks that way for now. I shall be experimenting with an SMT LM201 shortly, along with a few other changes for a forthcoming Rev 7. They way things are going, it’ll end up being entirely SMT eventually!

And then you can shrink it another 50%! ;-)

Or not....

[Kipling]

And then you can shrink it another 50%! ;-)

Or not....

I suppose it could be shrunk to eurorack size, but that would be quite a challenge!