ARP 3620 clone project

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: Kent, luketeaford, Joe.

Post Reply
User avatar
Kipling
Super Deluxe Wiggler
Posts: 1001
Joined: Tue Dec 10, 2013 3:31 pm
Location: South-East England

Re: ARP 3620 clone project

Post by Kipling » Mon Aug 03, 2020 3:34 pm

Good news Fernando

And well done to Mattia!

User avatar
solar1
Common Wiggler
Posts: 63
Joined: Fri Jul 31, 2015 5:14 pm
Location: USA

Re: ARP 3620 clone project

Post by solar1 » Tue Nov 17, 2020 3:23 pm

Hello Folks, I finished building the case for my 3620 clone well over a year ago, but finally thought I would post photos of it for those interested. As you can see, I used a 61 note keyboard assembly. The case is solid walnut with a plywood bottom. The front trim below the keyboard is detachable in order to allow slide-out removal of the keyboard chassis.
Attachments
100_2480.JPG
100_2457.JPG
100_2454.JPG
100_2474.JPG
100_2455.JPG

User avatar
Kipling
Super Deluxe Wiggler
Posts: 1001
Joined: Tue Dec 10, 2013 3:31 pm
Location: South-East England

Re: ARP 3620 clone project

Post by Kipling » Tue Nov 17, 2020 4:13 pm

Very nice work! :tu:

On another note I’m looking at this again and am wanting to do away with all the obsolete components by substitution where possible. I’m thinking 2N5458 instead of the 2N5459 N-channel JFET, and perhaps J175 instead of the 2N5461 P-channel JFET (though it has a different pinout). I know you can still get the original ones from China (and I have a bunch of each) but as to whether they are to the same spec or not, and whether it actually matters, I do not yet know.

What I really want to achieve is removing the two 2N3958 dual JFETs altogether, as I believe they are only creating a high impedance input stage to the LM1458 op-amp. By using a TL082 instead I’m wondering if I can simply bypass the JFETs but do I need to remove the pull-down resistors on each input as well? If not a change of value would probably be prudent as 100k may discharge the 470nF polycarbonate memory capacitors rather too quickly. If this isn’t going to work well then two 2N5458 back to back will have to suffice as I don’t think they need to be that well matched - I guess that a dual JFET was cheap and convenient in the 70s and FET op-amps either didn’t exist or were too expensive.

I’ll post some screen shots of existing vs proposed circuits ASAP but in the meantime has anyone any thoughts on this?

What I also fancy is developing a solution that derives the gate signal from the CV bus thus doing away with the need for a twin-bus keybed, though it would still have to be a mechanical contact keybed rather than bubble-contact diode matrix type (eg Fatar).

I have a small number of test PCBs (made to the original circuit design) I can experiment with so will update on progress.

User avatar
KSS
Super Deluxe Wiggler
Posts: 4800
Joined: Mon Jan 25, 2016 7:28 am

Re: ARP 3620 clone project

Post by KSS » Tue Nov 17, 2020 5:40 pm

Looking GOOD, solar1! :tu:

User avatar
KSS
Super Deluxe Wiggler
Posts: 4800
Joined: Mon Jan 25, 2016 7:28 am

Re: ARP 3620 clone project

Post by KSS » Tue Nov 17, 2020 5:48 pm

@Kipling
You definitely can remove the input FETS on the OPas. ;)
But use a good hi-Z OPA like a 412 or the 1013. Yes, the 100K resistors are removed. Member Broadwave has a thread here which discussed the details of how he did it. I've posted about it a few times when someone was having problems with their OG unit.

Dennis Colin has a patent on a single bus KBD which might be a useful path forward on eliminating the 2nd bus. Till.com lists the various ARP patents. Moog's too, where the 952 is a nice upgrade to the 3620's single trigger dual voice architecture. But that's probably a step too far for keeping this an ARP clone.

Edit: But the funny thing about that is that the moog dual voice keyboard uses ARPs patent! That's the reason there never was a lawsuit between them when ARP used moogs ladder filter -unpotted- in the 2500. They were BOTH in patent violation with the other!
Last edited by KSS on Tue Nov 17, 2020 6:07 pm, edited 1 time in total.

User avatar
Kipling
Super Deluxe Wiggler
Posts: 1001
Joined: Tue Dec 10, 2013 3:31 pm
Location: South-East England

Re: ARP 3620 clone project

Post by Kipling » Tue Nov 17, 2020 5:58 pm

KSS wrote:
Tue Nov 17, 2020 5:48 pm
@Kipling
You definitely can remove the input FETS on the OPas. ;)
But use a good hi-Z OPA like a 412 or the 1013. Yes, the 100K resistors are removed. Member Broadwave has a thread here which discussed the details of how he did it. I've posted about it a few times when someone was having problems with their OG unit.

Dennis Colin has a patent on a single bus KBD which might be a useful path forward on eliminating the 2nd bus. Till.com lists the various ARP patents. Moog's too, where the 952 is a nice upgrade to the 3620's single trigger dual voice architecture. But that's probably a step too far for keeping this an ARP clone.
Many thanks. I’ll look into this.

User avatar
solar1
Common Wiggler
Posts: 63
Joined: Fri Jul 31, 2015 5:14 pm
Location: USA

Re: ARP 3620 clone project

Post by solar1 » Fri Nov 20, 2020 1:15 pm

KSS wrote:
Tue Nov 17, 2020 5:40 pm
Looking GOOD, solar1! :tu:
thank you very much

by the way, for any folks interested, i think i still have an extra 61 note, multi-bus keyboard for this project.
PM' me if interested.

User avatar
Kipling
Super Deluxe Wiggler
Posts: 1001
Joined: Tue Dec 10, 2013 3:31 pm
Location: South-East England

Re: ARP 3620 clone project

Post by Kipling » Fri Nov 20, 2020 3:34 pm

I’m tempted by this, but it’s only a single bus Pratt Reed keybed. https://www.ebay.co.uk/itm/264941089274

I do have a Kimber Allen 61-note keybed that’s only single contact but their twin contact assemblies are still available.

Can you refresh my memory on how you managed to get the 61-note keybed to scale properly? I recall you made some resistor value adjusts. Cheers

User avatar
KSS
Super Deluxe Wiggler
Posts: 4800
Joined: Mon Jan 25, 2016 7:28 am

Re: ARP 3620 clone project

Post by KSS » Sat Nov 21, 2020 2:40 am

Kipling wrote:
Fri Nov 20, 2020 3:34 pm
I’m tempted by this, but it’s only a single bus Pratt Reed keybed.

I do have a Kimber Allen 61-note keybed that’s only single contact but their twin contact assemblies are still available.
I'd be more worried about the missing top C. Those are Hens teeth. Converting from single to multi-bus isn't too hard. And it's worth it.
I'll be addressing this as part of my new format and will post details. No timeframe for that except to say 'next year'.
for solar1 Kipling wrote: Can you refresh my memory on how you managed to get the 61-note keybed to scale properly? I recall you made some resistor value adjusts. Cheers
This is covered in the 2500 keyboard docs.That's the 3604 circuit, and the resistor values should be similar. If you still need it after solar1 replies, I can look into it for you.

Kimball organs from the 70s have P-R KBDs, and can often be had for free -in the USA. Those are most often single bus and 37 or 44 keys. 49 are *much* more rare and 61 even moreso. But they do come up and the big box is mostly empty on those with the 37 and 44 key models. 44 key models of the Kimballs usually have TWO KBDs. One of the two is often more well-suited to our needs than the other, but both can be usefully converted.
So that's one way to get that top C. Or the whole KBD. And an extra to boot.

Channel aftertouch can be added to most any P-R KBD. I much prefer a good P-R to the typical synth FATAR.

User avatar
Kipling
Super Deluxe Wiggler
Posts: 1001
Joined: Tue Dec 10, 2013 3:31 pm
Location: South-East England

Re: ARP 3620 clone project

Post by Kipling » Sat Nov 21, 2020 4:26 am

Thanks KSS

I think I've got a note somewhere of the changes required to use a 61-note keybed.

I'll pass on that 2500 style keybed - I probably wouldn't get around to using it anyway.

I will report on my progress regarding the modifications to remove the obsolete components as soon as I've built up a test board.

User avatar
Kipling
Super Deluxe Wiggler
Posts: 1001
Joined: Tue Dec 10, 2013 3:31 pm
Location: South-East England

Re: ARP 3620 clone project

Post by Kipling » Tue Dec 01, 2020 12:01 pm

I’m happy to report that my test board without the expensive obsolete parts works fine. OK it does still need good quality polycarbonate capacitors if possible but I have a bunch of them. For the record the 2N5459 is replaced by the still-available 2N5458, and the 2N5461 is replaced by a J175 (which has a different pinout). The 2N4392 is still available so remains in use as of course do the 2N3906’s.

Thanks go to KSS for his help on choosing a suitable FET OpAmp (LF412) and removal of the two input resistors.

I haven’t yet had a chance to look into the possibility of creating a gate signal from the CV bus but will give it some thought soon. I may also add a switch (probably pin headers and links) to enable a four or five octave keybed to be used - it’s quite a simple resistor substitution.

User avatar
KSS
Super Deluxe Wiggler
Posts: 4800
Joined: Mon Jan 25, 2016 7:28 am

Re: ARP 3620 clone project

Post by KSS » Tue Dec 01, 2020 8:07 pm

Sounds good!

PP or PPS will fill in fine for the PC caps. When your stock runs out.

User avatar
KSS
Super Deluxe Wiggler
Posts: 4800
Joined: Mon Jan 25, 2016 7:28 am

Re: ARP 3620 clone project

Post by KSS » Tue Dec 01, 2020 9:59 pm

Kipling wrote:
Tue Dec 01, 2020 12:01 pm
I haven’t yet had a chance to look into the possibility of creating a gate signal from the CV bus but will give it some thought soon.
What type of gate signal are you after? I've tried a few different ways to get there, if you mean a second gate to go with the duophonic CV.
Edit: I think you may mean deriving gate from sngle bus KBD contacts? For that, there's the Colin patent already mentioned? And moog had a few 1 bus circuits too. /edit
I may also add a switch (probably pin headers and links) to enable a four or five octave keybed to be used - it’s quite a simple resistor substitution.
I would suggest including all the normal widths? 25,37,44,49,61. As you say it's only a simple resistor change.
Attached is a cartoon-ish graphic of my reasons for delving into these issues. It's planned to work with small to large span keybeds too, and you can see the hole for the input cal switch just under the A of the logo upper left. I'm still not sure if this will be a pot or a PCB rotary switch in the final rev. You may not need the same panel access. This is intended to interface with any resistance and gate circuit. Like a ribbon, or a big arm pot like W Carlos pitch select circon thingie. Those are what the input DIN is there for. Shown 'covered' here.

You can also see the dual gate and trigs for the aux voice.

Sharing this here also to say that since this 1049 is part of my new 50SQ format, I've been working on some of the things Kipling's project builders can use too. Like the drawings and resources for making an accurate 3620 clone at various widths. My 50SQ project is eventually going to be fully open hardware, and the mechanical parts will be shared with builders of this project too. As before, no timeframe except next year. So build Kiplings PCB-Panel and know that you'll have answers for the keybed, contacts and other possible issues too in the new year.
In respect of what Kipling has done -in both this and his 1601 project- i'll be opening the KBD part of the new format sooner than some of its other parts. We need more DIY kbds and controllers! And if they're ARP derived? So much the better.
Attachments
APT-50SQ-1049-4a-CLR.JPG
Treat utility modules as stars instead of backup singers.
Treat power supplies like Rockstars instead of roadies!
Chase magic sound, not magic parts.

User avatar
fallout
Common Wiggler
Posts: 83
Joined: Fri Dec 18, 2020 2:45 pm

Re: ARP 3620 clone project

Post by fallout » Thu Dec 24, 2020 10:26 am

Question. I have a 49 note keybed from a junked Lowrey organ that I'm considering using for this project. I would have to retrofit some dual contact key switches. Are there any other sources for these type of key contacts besides Kimber-Allen or any other options?

Thanks!

User avatar
KSS
Super Deluxe Wiggler
Posts: 4800
Joined: Mon Jan 25, 2016 7:28 am

Re: ARP 3620 clone project

Post by KSS » Thu Dec 24, 2020 8:49 pm

Welcome to MuffWiggler, fallout!

Use optima gold 014 guitar string-wire and make your own KBD switches using stripboard or a DIY PCB. Much better life and action than the leaf switches, inexpensive and repairable.

I wrote more details about how to do so in some thread here, but will be expanding on that in the coming year as my post above said.

An alternative approach is to use reed switches. Expensive per switch, but simple to fit into weird KBD configs and have excellent life and switching qualities.

If you post a photo of your Lowrey KBD -waterfall type?- I might be able to give you specific advice for it. I have a few Lowrey KBDs here I've been saving for Roland clones, and if similar to yours may aid in pulling dimensions and methods for getting you some good results with yours.

User avatar
LED-man
Super Deluxe Wiggler
Posts: 1528
Joined: Wed Aug 08, 2012 1:48 pm
Location: Wolfsburg Germany
Contact:

Re: ARP 3620 clone project

Post by LED-man » Fri Dec 25, 2020 8:49 am

The Ms2600KBD use a Fatar keybed.
https://diy-mms.jimdo.com/ms-2600-kbd/
################################
TTSH Support Page and other DIY Projects:
https://www.DSL-man.de

User avatar
fallout
Common Wiggler
Posts: 83
Joined: Fri Dec 18, 2020 2:45 pm

Re: ARP 3620 clone project

Post by fallout » Fri Dec 25, 2020 9:59 am

KSS wrote:
Thu Dec 24, 2020 8:49 pm
Welcome to MuffWiggler, fallout!

Use optima gold 014 guitar string-wire and make your own KBD switches using stripboard or a DIY PCB. Much better life and action than the leaf switches, inexpensive and repairable.

I wrote more details about how to do so in some thread here, but will be expanding on that in the coming year as my post above said.

An alternative approach is to use reed switches. Expensive per switch, but simple to fit into weird KBD configs and have excellent life and switching qualities.

If you post a photo of your Lowrey KBD -waterfall type?- I might be able to give you specific advice for it. I have a few Lowrey KBDs here I've been saving for Roland clones, and if similar to yours may aid in pulling dimensions and methods for getting you some good results with yours.
Hi and thank you! I'll take a pic of the keybed this weekend. I have an idea on how to make it work but certainly open to suggestions or better ideas. I also managed to grab a bunch of Kimber Allen key contacts from the bay. They're the wrong size (4 contacts) but I'd imagine it will be fairly easy to disassemble them to make the correct 2 contact switches. I'll just have to figure out how to rivet them back together.
LED-man wrote:
Fri Dec 25, 2020 8:49 am
The Ms2600KBD use a Fatar keybed.
https://diy-mms.jimdo.com/ms-2600-kbd/
That looks awesome! I'm also considering building one with a Fatar keybed and decoder. I'm still wrapping my head around the whole project.

User avatar
fallout
Common Wiggler
Posts: 83
Joined: Fri Dec 18, 2020 2:45 pm

Re: ARP 3620 clone project

Post by fallout » Fri Dec 25, 2020 11:35 am

BTW the MS2600NG itself looks excellent.

User avatar
Kipling
Super Deluxe Wiggler
Posts: 1001
Joined: Tue Dec 10, 2013 3:31 pm
Location: South-East England

Re: ARP 3620 clone project

Post by Kipling » Tue Dec 29, 2020 8:35 am

LED-man wrote:
Fri Dec 25, 2020 8:49 am
The Ms2600KBD use a Fatar keybed.
https://diy-mms.jimdo.com/ms-2600-kbd/
Hi Patrick

Is this project going to be available to buy at some point? Using Google Translate I can't find any reference to how to buy a kit or the PCBs etc.

User avatar
LED-man
Super Deluxe Wiggler
Posts: 1528
Joined: Wed Aug 08, 2012 1:48 pm
Location: Wolfsburg Germany
Contact:

Re: ARP 3620 clone project

Post by LED-man » Tue Dec 29, 2020 4:26 pm

He made one sale for diy fullkits, 700€ Including everything except Tolex and the special tolex glue.
The plan was that I offer the pcbs after his first sale.

But I’m sure we can use a fatar keybed with an addional microcontroller with the arp3620 Controller too ?
Tauntek made a pcb for his minimoog clone too, which scans a diode based keybed and creates a 1v/oct and trigger.
################################
TTSH Support Page and other DIY Projects:
https://www.DSL-man.de

RadekTymecki
Common Wiggler
Posts: 68
Joined: Fri Apr 12, 2013 4:38 am

Re: ARP 3620 clone project

Post by RadekTymecki » Tue Dec 29, 2020 7:09 pm

Kipling wrote:
Tue Dec 29, 2020 8:35 am
LED-man wrote:
Fri Dec 25, 2020 8:49 am
The Ms2600KBD use a Fatar keybed.
https://diy-mms.jimdo.com/ms-2600-kbd/
Hi Patrick

Is this project going to be available to buy at some point? Using Google Translate I can't find any reference to how to buy a kit or the PCBs etc.
I am not sure. As for MS2600NG I am the one who got the last kit... I do not know about the keyboard. I have sn 030 and I do not know if there are going to be more. You can ask Matthias via contact form on his website

User avatar
fallout
Common Wiggler
Posts: 83
Joined: Fri Dec 18, 2020 2:45 pm

Re: ARP 3620 clone project

Post by fallout » Tue Dec 29, 2020 7:18 pm

I would certainly be interested if the MS2600NG was going to be made available as well.

User avatar
fallout
Common Wiggler
Posts: 83
Joined: Fri Dec 18, 2020 2:45 pm

Re: ARP 3620 clone project

Post by fallout » Mon Jan 04, 2021 7:59 pm

Here's a few pics of the underside of the Lowrey organ keybed that I would like to try and use. It seems like I have 2 problems to overcome. First problem is that the underside of keys do not clear the of the keybed frame. The keys would not make contact with the Kimber Allen switches when depressed. As you can see in the pictures, I should be able to put something in the keys that would make contact with the key contacts. I open to suggestions of what I can use or any other creative suggestions. I'm thinking of some kind of plastic 'tee' but still not sure what I can use.

The second problem is that the Kimber Allen switches are too big and wind up hitting the frame. My idea was to pop out the rear rivets/dowels and hacksaw the switches into a usable size. I would also like to remove the unused contacts as well to make it easier to solder the key and gate buses.

Thoughts? Suggestions? Thanks!
Attachments
IMG_6689.jpg
IMG_6687.jpeg

User avatar
fallout
Common Wiggler
Posts: 83
Joined: Fri Dec 18, 2020 2:45 pm

Re: ARP 3620 clone project

Post by fallout » Mon Jan 04, 2021 8:55 pm

I tried disassembling one of the KA switches and I don't think my idea of cutting them to size will work or at least not without a lot of work. The effort required wouldn't be worth the trouble. The little black dowels on the rear of the switch hold the wires in place and without them, the contact wires move freely. Back to the drawing board!

User avatar
KSS
Super Deluxe Wiggler
Posts: 4800
Joined: Mon Jan 25, 2016 7:28 am

Re: ARP 3620 clone project

Post by KSS » Mon Jan 04, 2021 9:35 pm

What level of mechanical DIY skills and tools do you have?
My first early DIY synth KBDs in the 70s used military surplus leaf switches which were dual screw mounted in stacks like these. Removing the rivets of yours and using screws instead should be simple. I'd use a separate piece of extruded aluminum tapped for re-mounting them. And I'd NOT cut the switches! Angle them instead to fit and you only need one end of the KBD frame to extend enough to accomodate this. Actually if you use the suggested extruded switch mounting plate, only *it* needs extend beyond the key width and you can probably 'hide' this relatively small 'overhang' in a left or right end cover which is pretty typical in KBDs anyways. A 3620 has 3/4" available either side for example. Just cut away the pine tolex covered piece to clear your switch and its carrier.

To extend the key for actuating the switches, the most flexible and configurable option would be to design something and use a 3D printer or service to make them.
Absent that option, something as simple as properly cut-to-sized rectangles from plastic sheet nylon, PE, HDPE, UHMW, Acetal-Delrin would all be suitable. The three 'vanes' extending into each key's underside actuating rectangle suggest a round rod of any compliant type plastics listed above would work well. But its end being round doesn't work well with your switch requirements. You'll need some type of 'blade' with a 'line' contact for those anyways. Which means the actuator transition piece below the 'blade' and the blade itelf could be different materials or pieces and joined.

Using the separate switch mounting plate gives you more options and adjustability; simpler repair and initial switch calibration. Along with not needing to work much with the existing steel KBD frame. You can probably find existing mounting holes to use, and if you look at how the Pratt-Read KBDs mount their j-wires it might help in your design.

Which also brings up the possibility of using a PCB -DIY or sent out for build- instead of the aluminum mounting plate. You could make a one octave PCB and use it four times. This could also help immensely in adding the resistors and caps of each key if you use the original electrical design. don';t forget the 49th key!

Finally, once you've gone that far, it's only a very short distance from there to use the wires themselves individually on the PCB and forget the rest. Or to use the gold guitar wire shaped to become the needed switches. Still might use the switches you have for the first one, but having the PCb already designed and usable -with mods to a second rev- for others following after your lead.

Edit: this reply was written before the post just above it was made. Glad you see that cutting the switches is a no-go.
Treat utility modules as stars instead of backup singers.
Treat power supplies like Rockstars instead of roadies!
Chase magic sound, not magic parts.

Post Reply

Return to “Music Tech DIY”