Roland AP-7 Jet Phaser - let's get it working! (Aion FX Octatron)

aion

aion

Authorized Vendor
(Adapted from this thread)

About four years ago I took a shot at the Roland Jet Phaser:

Roland AP-7 Jet Phaser / Aion FX Octatron

The first prototype sounded great. I was about a week away from releasing it when I decided to build a second prototype just to be extra thorough... and with no obvious differences in the build, this one had the dreaded "thunk" at the bottom of the sweep that has plagued pretty much all DIY attempts over the past two decades, e.g.:


If I biased it so the "thunk" was gone, then the phase was too weak and it didn't sound anything like the original.

It got bumped from the release schedule until I could spend more time with it, and somehow four years have gone by. I may be able to get to the bottom of it eventually, but it's been a very hectic year with (waves hands in the general direction of Washington, D.C.) and I can't give it much direct focus for the time being, certainly not for the next few months.

If anyone wants to take a stab at it and discuss their findings in this thread, I have PCBs that I would be happy to send out for free. I have a full build doc, schematic, etc. as well as JFET data measured from four different vintage units (I managed to accumulate three of my own during the development process). I can also supply matched octets of 2SK208-Y JFETs on SMD adapters along with the boards, so it cuts out the most annoying part of the build.

I dug deep into the archives on this one and I don't think I found any reports of a successful DIY build of this circuit, and so it was a bit Ozymandiesque of me to think I could tackle it... but if we can figure out on a technical level where the thunk comes from, and we crack the circuit in a reliable, repeatable, and public way, then the whole DIY community benefits.

Note: I have around 12 prototype boards available and 8 are already claimed, so I'll be a bit more selective on the last four, but if you have a strong electronics background (particularly theory) and a good understanding of the technical concepts of this circuit, and a willingness to volunteer some time for Science™, then send me a PM.

I'll follow up with all the research I've gathered on this circuit so we can all start from the same place.
 

Attachments

  • jet-phaser.png
    jet-phaser.png
    147.3 KB · Views: 21
Here's the build doc:

Much of the info I have is compiled there between the design notes and build notes, but I do have some other source documents and notes that I'll post as well - stuff that was too in-depth for the document itself, or things that informed the design process but wouldn't be as relevant to someone building it.

Also - please don't post the PDF anywhere else - not that it's secret or anything, I just don't want it to get indexed by search engines and have it turn up when people search for "octatron build pdf" after the project is released and they end up using the wrong info. I have it blocked from crawlers on my domain, but wouldn't have the same control if it was posted elsewhere.
 
God speed everyone. I'm a caveman but there are some serious nerds here that should be able to make progress. I'd definitely be in line to but once its ready.
 
Have you tried another opamp for the phase section?

From the descriptions of the issue, it sounds like the lfo bottom is hitting a limit. The question is then... the limit of what?
 
Believe it or not, but the Roland Jet Phaser was my very first guitar pedal. Such a wonderful and crazy pedal. Still have it and appreciate it more now, than I did back then ;-)
 
going to set aside some time this morning to look at the WA I have sitting around and check some of the notes from the various threads with Hammer and Keen because I'm betting WA played with resistors and caps rather than hitting a very specific bucket with the transistors to get past the bottoming out problem
 
ok so part numbers will reference the aion schematic
"Just ahead of the Feedback pot you'll see a 10uf cap. Perhaps this might be a bit high and noit trim back on the bass peaks enough. Consider droping its value to something under 1uf." - Mark Hammer, referencing the AP7 service manual part value https://www.diystompboxes.com/smfforum/index.php?topic=97437.0 .

both WA and Aion have this cap as 1uf (c31)

"The other thing of note is that the grainy Roland pdf schematic shows a 330k resistor in the feedback loop of the very last phase-shift stage. Since this adds gain, it places the feedback signal at risk of exceeding the threshold of "ugly clipping" for the FETs. Maybe drop that down to 220k or perhaps even make it uniform with the other stages at 100k." - Mark Hammer, same diysb thread quoted above

WA has this part specifically silkscreened with the value 330K (r62) bold and underlined, don't specifically know what that means but my guesses are production error where they made all the stages match OR they tried this advice and it didn't work.

I do notice the aion build relies on a charge pump and I do wonder what the results look like using 18v power supply instead
 
ICTRock said:
ok so part numbers will reference the aion schematic
"Just ahead of the Feedback pot you'll see a 10uf cap. Perhaps this might be a bit high and noit trim back on the bass peaks enough. Consider droping its value to something under 1uf." - Mark Hammer, referencing the AP7 service manual part value https://www.diystompboxes.com/smfforum/index.php?topic=97437.0 .

both WA and Aion have this cap as 1uf (c31)
Click to expand...

Yeah, this was a change between the 1975 version and the '76 revision. Roland dropped it to 1uF, and in some cases two 1uF in series for 0.5uF. Most of the DIY attempts have been based on the '76 schematic... it could be that reducing it further might help, and is something I would definitely try, but the issue does still happen with 1uF.

ICTRock said:
"The other thing of note is that the grainy Roland pdf schematic shows a 330k resistor in the feedback loop of the very last phase-shift stage. Since this adds gain, it places the feedback signal at risk of exceeding the threshold of "ugly clipping" for the FETs. Maybe drop that down to 220k or perhaps even make it uniform with the other stages at 100k." - Mark Hammer, same diysb thread quoted above

WA has this part specifically silkscreened with the value 330K (r62) bold and underlined, don't specifically know what that means but my guesses are production error where they made all the stages match OR they tried this advice and it didn't work.
Click to expand...

Another '75-'76 change - it was originally 100k in the '75 version but they bumped it up to 330k. I would have thought the same, that it would be better with less resistance (less gain from the phase stages) but I believe I tried it both ways with no change.

ICTRock said:
I do notice the aion build relies on a charge pump and I do wonder what the results look like using 18v power supply instead
Click to expand...

It's +/-9V, so the positive rail is straight from the supply and only the negative rail is coming from the charge pump. You would have to do two batteries in series like the original to A/B test.
 
quick notes for those with the WA ... the six tiny screws on the back and the nut on the rotary switch are how you disassemble it. maybe that will save some of you some steps
 
I have started on the template but I am waiting to hear back on a few in consistencies. I am sure he threw this together from a hodgepodge of build docs and I want to verify some of the numbers. As of now the numbers don’t add up, lol. I am using my neurotron(verified) as a guide and messed it up on Tayda so hopefully no one needs it until I fix it, lol.
 
shouldn't, it's part of the fuzz circuit. fwiw the WA is all LM324, I expect the roland varied depending on supply. the service manual references two different quad op amps
 
Has anyone considered emailing Warm Audio to see if they know how to fix it? I wouldn't expect much necessarily because this is sort of competition for them, indirectly, if you squint a little, but it's worth a shot.
 
ICTRock said:
shouldn't, it's part of the fuzz circuit. fwiw the WA is all LM324, I expect the roland varied depending on supply. the service manual references two different quad op amps
Click to expand...
Does the WA LM324 have a B or BA suffix? I think the BA version has a slightly lower offset voltage
 
You must log in or register to reply here.
Back
Top