My Mini Moog Voyager. Click to go to my web site...

Tuesday, June 29, 2010

Change in plan

So, the total knee has given me a lot of time to sit and at least try to think. It occurred to me that I really need to consider this project not as a "one-off" system for me. With Nyle's encouragement to share and release this stuff,  I realized (finally...) I need to take a more universal approach to the PCB/panel designs and consider those who are going to stuff just one or two of the modules into an existing Doepfer rack or something. I've been looking at the Doepfer racks and they can be as shallow as 80mm in front of the power supply itself. It looks like the general minimum is about 100mm and I think most modules will fit this form factor. The sequencer as it stands now may be a problem, but since the panel will be so wide, I can rethink it and make it wider, but shallower.

I also decided I'd better put a 10 pin 0.1 inch spaced power header for the true euro rack users. I'm keeping the 4 pin, .156 inch spaced connectors for me and the MOTM/Frac guys. I am seriously considering throwing a 10 volt regulator on each PCB, too. But, that seems like a hack to me somehow. When I get the next set of prototypes built, I'll just test them on +/- 12 and if they work, I'm not going to worry about it. If they don't, I might just have to put the regulators on the PCBs since I doubt many people will dedicate a -10 volt power supply for 1 or 2 modules and I'd really like to see people build these.

So, here's how the plan is changed:

  1. Pick a "standard" size board and keep the depth to less than 100 mm if at all possible (the sequencer and VCO may be a problem). I'll stick to 3.95 inch tall (just under 100 mm euro rack standard) so they will fit vertically in a Blacet rack or euro rack.
  2. I'm not going to do PCB mounted pots but will have headers to facilitate using MTA connectors and flying wires.
  3. I will pick a pot spacing for the panels I design to accommodate a spacing compatible with a knob like the Alco PKE60B1/4 or PKES60B1/4 and an Alpha 16 mm pot. The knobs have a retro look, but are smaller than the original Steiner knobs which will let me shrink the module widths for the EGs and VCOs at least. They aren't much bigger than the soft touch knobs Blacet uses or the Doepfer knobs, but I think they look more in keeping with the original. A spacing of 1" maybe 1.1", for the pots should be good. I know that you can get as close a 0.75 inches on a Frac or Euro panel using smaller knobs, but it's really tight. I don't want to waste space, but I need some room for these chunky fingers of mine. Staying away from PCB mounted pots also gives the user a much wider choice of parts and mounting configurations.
  4. I will need to design a custom module and/or daughter board which will interface the S-Triggers and Voltage Triggers. I'm not going to re-work the Steiner modules for this as they were designed to take advantage of the way S-Triggers can be connected together in an "OR" gate fashion. Some users may not care and will need to use voltage triggers to trigger the EGs, or sequencer. So, I will design and build some boards to do this as separate modules.
  5. I'm going to look into the advantage of just having a 10 volt regulator on board. It takes space and a few more components, but if the modules need it, this may really be the only way to do it efficiently. For my part, if I need -10 volts, I'll just change the Power-One supplies to give me +12 and -10.
  6. Add a Doepfer header to the board for the Euro guys and leave the 4 pin MOTM for the rest of us.

Sunday, June 27, 2010

Logo and panel designs

Still working on parts lists and waiting on the delivery of prototype PCBs. I want to get some of the core modules tested working and ready to go before I do too much more capture and layout. Too many irons in the fire, etc. etc. etc...

I'll be sending in an order to Mouser this week with production parts for the VCO, VCF, VCA, and EGs, maybe the sequencer, too. I'm going to try a replacement Germanium diode and another UJT in the sequencer to make sure there are some readily available parts for those who are wanting to build this stuff.

I have a good first go at the EG panel and the Power Panel. The power panel will take up the full rear of the Euro rack and is designed to take the Power One HAA15-0.8 supply, a combination plug/fuse/switch assembly, and some power distribution boards which follow John Blacet's dimensions for his distribution boards. I laid out my own version which will supply 10 modules which I will also publish. The problem is I did it in Target 3000! and not in Eagle, so it won't be as universal. I allowed for two boards because there was room.

I still need to print out the EG panel and test fit parts, but based on other panels I've done, I'm pretty sure it will work out. I'm going to use alpha 16mm pots with round shafts which should give some good knob options. I decided to not use the "Blacet" style soft touch knobs and go with something more true to the original Steiner systems. I allowed 1.33 inches between knobs horizontally and 1 inch vertically so there should be room for big fingers like mine. I've found you can go to 0.75 inch centers with small knobs, so 1 inch should be luxurious.

I'm using the Schaffer/Front Panel Express software and will be having them do panels for me. I'll make the .FPD files available as they are proofed. Understanding this is a DIY project, I'd still like to work out the bugs before I start releasing the final product.

Last note today. I had a total knee replacement on Monday, June 21. An old injury finally got the best of the knee but I'm healing. I'm planning on a meeting with Nyle mid to end of July and should have the rest of what I need to get information out there.

As always, I appreciate comments and suggestions.

Stay tuned.

Friday, June 25, 2010

VCF schematic

The first of the schematics. This is the production version of the Synthasystem VCF. There are some subtle differences from the other Steiner VCF clones, notably  in the feedback loop and "Q" and in the variable resistor "diode" chain. The PCB I laid out makes it possible to use discrete transistors or a match pair like a SSM2210. Nyle had a look at this one a said it looked like it incorporated the latest revisions.


Sunday, June 20, 2010

Power supplies

I think I posted I have the rack frames and the Power-One power supplies. I am a little at a loss, however, regarding how to best mount the supplies, run the AC, and run the distribution board. The PS is deep enough that it will need to have its own several inches of horizontal rack space. I think I will mount it on as narrow a panel as I can with a power switch and LEDs and just live with it. I will be able to mount it on the front panel or rear. I just might have to put a blank panel on the front the same width as the PS I'll have to actually lay out the panel and build one to see I think.

I've also been working on parts lists. Eagle has a ULP which makes the work pretty easy, but they still take a bit of hand editing to group the parts exactly as I want them. I also added some notes. I think I will try to get an official order for parts off this week at least for the VCOs, VCAs, Sequencer, and ENVDLs. I'm hoping the next batch of boards will be the production boards for these modules. I'm trying to be patient regarding the filter. I still want to lay it out following Nyle's original artwork.

The last batch of boards still hasn't entered the production queue at PCB-POOL which is unusual.

The previous order including the Sequencer should be here any day and I should be able to get the important bits for the sequencer soldered up and tested. I'm mostly interested in the reset functions to make sure I read the schematic properly.

I've sent the schematics I finished off to Nyle and he said he was going to look them over to make sure I had the latest revisions which is great! When I visit with him later in July, I'm going to print off the electronic copies I have so I can make some notes as he and I talk.

Not a lot new, but still, making progress. It's going about as fast as I thought it would. 

If anyone is actually reading this, as I have working PCBs, I will post here and in a couple of the forums and SDIY list with information about getting the schematics, PCB layouts and/or PCBs themselves. I will also make some pages on my web site specifically for this project with build notes, photos, etc.

Thursday, June 17, 2010

VCO prototype PCB

Here is a view of the current VCO PCB. It uses the most recent VCO core design (circa 1976) and the shapers from the original Synthasystem. I've laid it out so you can use hand matched transistor pairs in the exponential converter, an SSM2210 or LS319 and a discrete PNP, or a THAT340 which has a pair of matched NPNs and PNPs (only one PNP is needed). The exponential pair needs to be in thermal contact with the PNP and thermistor and I thought the THAT chip would be a great choice. I've tested a THAT chip in the core and it worked fine.

I've included a 4 pin MTA connector ala MOTM and FRAC and some simple signal conditioning which was not present on the original. It can certainly be bypassed to have a "true" Steiner VCO. The input is marked +12 and -10 but I'm pretty sure it can be run on +/- 12 VDC. I've not tried it yet as I'm waiting for this "next" version PCB to arrive.

I have ground pads on each pot and jack rather than a star grounding system. I just haven't found this to be a problem and when I look at others who know much more than me, the layouts tend to be like this. I also tried to give enough room to use 0.1 inch center connectors for the pots, jacks, etc. I chose to not use PCB mounted pots so you can put it behind any panel you want.

I'm not posting the official artwork or schematic yet because I want test it all first and go over it with Nyle.

So enjoy, but this is not the final version, at least not yet and I can't guarantee it will work.

I'm happy to entertain suggestions on the layout.


A couple of comments on JFETS. Nyle told me he had problems with the 2N5163s. That may explain the problem with the NOS JFETS I bought on eBay and the two "dead" JFETS from my old stock. All the more reason to stick with the 2N5246s. He also said pretty much any n-channel JFET would work here.

I also gt some insight into why he used so few ICs. He said he was worried about obsolescence and not being able to find an IC down the road. A man after my own heart :) As the schematics are put up, you will see there are very few ICs beyond a '741 OpAmp here and there.

Parts coming in

I received the multiplier chips for the ring modulator, the 3140 opamps for the duophonic keyboard, power supplies (now to figure out how to re-calibrate them to -10VDC), and the Euro Rack frames. I've ordered a second go around for the VCO and VCA PCBs, and 10 envelope generator PCBs (the ENVDL type). I need to order some more resistors. I didn't have enough of some values to finish the first VCO and had to jury-rig some.

I'm expecting the Sequencer PCB soon and as soon as I've verified some of the switch operations, I'll order the first production run of 3 or 4 boards.

I've been working on some panel layouts, too. The VCO is well under way, as is the Tuner/Monitor (I am adding room for a mini-scope) and triple envelope generator.

Monday, I have total knee surgery scheduled, so I'm hoping after the first few days, I'll feel good enough to get more schematics captured and board layouts done.

Wednesday, June 16, 2010

Update on distribution of schematics, etc.

I had another long talk with Nyle today and he is fine with sharing the schematics and PCBs. So, I'll think about the best way to do it, and will share as I can.

Keep watching here for information on how this is going to happen.

Sunday, June 13, 2010

No Pain, No Gain, seriously, no gain...

So, the envelope generator works flawlessly. I was instantly in love with the damp and quathin switches. I have to find out where the term quathin came from.

But, the VCO is a different matter. I didn't want to risk destruction of one of my original 2N5163s, so I used a NOS which I bought on Ebay. Turned it on and no oscillating. I removed the NOS 5163 and installed an original one. The saw core started to work fine. The frequency adjust takes it from below hearing to above and all seems good. I then decided to try all my old stock and found two JFETs which had died. None of the NOS worked, very disappointed. I then tried the 2N5246 JFET Nyle spec'd in his later VCO core and they worked like a charm. I tried the NTE312s, and NTE132s, and they worked great. I then tried the 2N5246s which did not work. So, I think I'll probably just go with the 2N5246s.  They are in production, and inexpensive. To that end, I changed the schematics to reflect the 5246 footprint as it is different than the 5163.

The bad news is that the saw core output doesn't make it through any of the shapers. I really don't understand why. I have checked the original schematic against mine several times now and I don't see any errors. I've eyeballed it logically and the connections appear to make sense as far as I can understand them. I double checked the pinouts on the Eagle models against the data sheets for the SSM2210s, THAT340, and the transistors. I even put my transistor analyzer on the transistors to verify which pins are which and checked that against the PCB footprints, and how I installed everything. So, I get a nice sawtooth out of the core, but no gain, no shaping, no output.

I'm out of town for a couple of days, but when I get back I'll have another look. I'm not too worried because I've changed a lot for the next revision, so I will probably wait til I get the next prototype ordered and stuffed.

I've also made some decisions regarding the schematic capture and PCB design. I am doing the schematics exactly as Nyle had them electrically. For example, 2 meg pots are hard to find, but I'm entering them anyway with the associated capacitor values. I didn't do that to start, but I think it is important to get them entered the way Nyle had them for the sake of history and preservation. Later, when I build one or two for my other modular,  I might make some "improvements". Use OpAmp buffered outputs/inputs, etc. But for now, I want this thing to be as close to what I would have had in 1975 as possible.

For little things, I will make value adjustments accordingly as I build and order parts. Like, there are several places where 2 Meg audio pots are used. 2 Meg pots are not easy to find. I'll use 1 Meg and then up the capacitor to get the same time constant, but the board and schematic will reflect the 2 Meg pot. It will also reflect the 2N5172s and the in production PN5138s (different case from the 2N5138 and slightly different footprint, straight line instead of a semi circle). The pins will be marked so anyone can use the NPN/PNPs they want with relative ease.

 That said, I am using footprints which will allow newer parts, like NPN matched pair ICs to be used i f wanted. I confirmed the THAT340 NPN/PNP IC works fine in the exponential convertor and for my personal production, I'm going to use them. I'm trying to include copious notes on the PCBs to help in construction and am labeling the PCBs with component values instead of component number. I am also including any notes, etc. Nyle made on his schematics to preserve the archive.

Last, I got hold of Nyle by phone last week. I was as  nervous as I was that first time I met him in 1975. Weird how the psychology of that works. Anyway, we chatted for a few minutes and decided we should meet and talk about what I'm doing. As Nyle put it, he likes to be involved with people doing fun things. I was flattered he thought what I was doing with his designs was fun. Anyway, I'm hopeful now after talking to Nyle these unique designs will be shared in some form. I don't know what, how, when, or where, but I have my fingers crossed.

Stay tuned...

Friday, June 11, 2010

More progress

I received the VCO prototype boards today. They are the 1974 vintage. I really hope tomorrow I'll be able to make time to put a VCO and/or the EG together and try them out. I also received the shipment of new JFETS and I'll try one in the VCO to verify that it works. I also received confirmation the Sequencer, VCF, and VCA prototype boards have shipped. I don't know if I'll build the VCF as it is different than the one in the Synthasystem, but I'll at least get part of the sequencer built to test the clock and stepping functions, even if I don't get all the pots, etc. wired up

I received the Power One power supplies today so I'll need to figure out how to change them to +12/-10, which should be doable.

The best news is I spoke to Nyle on the phone and I'm going to try to meet with him next week. It was great to re-connect with him and catch up. He said he was excited to see the PCB layouts and maybe do something with them. As I get more details on that note, I'll post here at the least.

Monday, June 7, 2010


I decided I really needed to build a Steiner keyboard. He offered a 3 and 4 octave at least from the owner's manual. I found a new "old stock" Pratt-Read keyboard from Stephen at Synthwood! That means I won't have to hack a newer one, and it will be truly vintage. I also ordered a set of mod wheels from Synthwood. I'll use the one wheel for the pitch bend. Don't know yet what the other wheel will be good for. Maybe the portamento.

I also ran down some obscure parts for the ring modulator and the keyboard uses a odd MOSFET OpAmp. But, I found them all and have ordered some and to spare. I think the last hard part to find will be some 0.01% 22k resistors. Precision Resistor should be able to make them for me. I'll have to look into it.

Saturday, June 5, 2010

More Progress

Wow, a late night. The ENVDL boards came yesterday so I'm hoping to find time to stuff one and try it out. I also spent a lot of time updating some schematics/PCBs. It turns out the resonance and output of the Synthasystem VCF is different than the Synthacon, or even Nyle's 1974 paper. There are some "penciled" modifications which are a bit confusing, but I think I got it figured out.

The VCO information included an updated VCO core, 1976, which uses a different JFET and some other small differences. I re-did the VCO Function Generator based on the newer VCO MOD schematic. I think I'll build them both in prototype to see if one is any better. The nice thing about the newer one is the JFET is still in production, 2N5246, as opposed to the obsolete 2N5163. It seems to be a pretty good cross reference and I will try them in the other circuits like the sample and hold, etc. Since Nyle used them, and spec'd them, I can still keep the circuits original.

After I had looked through the schematics, it seems there are two which are missing. If Nyle doesn't have them, I can only assume I'll never have them. The missing two were the Voltage Follower, basically a lag, or portamento circuit, and the Voltage processor, a module which simply changes and recombines the phases of the input signals.

For the VFollower, I think I can just pull the portamento part out of one of the keyboard schematics. If I understand it right, it's little more than an adjustable low pass filter. The intended use was for control voltages like from the sequencer I think.

For the VProcessor, I'm sure it wasn't much more than some 741 OpAmps configured in normal and inverting stages. I think even I can design something like this. I have a pretty good feel for the general design practices Nyle used and so will use comparable resistors, etc. to put it together. I want to get some of the other modules laid out first though.

It's too bad I sent in the VCF design I had before I had laid out the Synthacon version. The VCO is fine as it uses the original Synthasystem schematic so I can test that core, etc. The sequencer, too should be fine.

I've optimized all of them since then anyway, so I will need to resubmit the lot. 

So, next up will be to build an ENVDL type 1 and 2 and try them out. Continue to capture the schematics, and layout PCBs.

Friday, June 4, 2010

Big news

The big news today is I re-connected with Nyle and I now have all the schematics I need. Just to be clear, I can't share these. I'll post photos, sounds, but nothing that can really be used to reverse engineer the modules, I'm sorry. I owe Nyle that respect and courtesy.

So, now on towards the design! There are some subtle differences in the VCO cores and so I'll have to decide which one to implement. I think the latest one will be where I go, especially since it uses a still-in-production JFET as opposed to the 2N5163 in the earlier design.

Looking at the other schematics, the only other part which seems obsolete is the UJT, but I'm satisfied pretty much any UJT will do the trick. So, not only are Nyle's designs very unique, they are still current and I think I can build them all with in production parts and make them essentially identical to the original.

I'm especially excited for some of the utility modules. I was going to settle for some contemporary designs, like for the tuner/headphone amp, but now I can make the original deal.

I think the biggest project in this will be the keyboard. There are three keyboards, the Synthacon, a two voice Synthasystem, and a "multiphonic" keyboard. At the time, I only had the two voice Synthasystem and so will probably do that one. I'll go ahead and build the keyboard electronics absent a keyboard and will decide what to do about the keyboard later. I'll need to find an old Pratt and Reed or some other buss type keyboard.

Later for this though. I can play the thing with any number of voltage sources. As long as I get a CV and gate, I can have fun. I will need to build one module that will convert voltage gates to S-type triggers/gates, but that should be easy.

Tuesday, June 1, 2010

More PCBs done!

I finished the first go around for the VCA, VCF, and Studio  Sequencer PCBs. I was most hesitant on the Sequencer as its schematic was the most problematic to read, in particular with regard to some of the off board signals. But, I had an epiphany and it seemed to fit together. I ordered them today and should have them in 2 or 3 weeks. Just in time for my total knee replacement...

The VCO boards have now shipped, so during the next month or so, I should actually be able to get a core set of modules together to make some Steiner-Parker sounds! When I do, I'll figure out how to post them here or on my web site.

The sequencer board is really much bigger and sparse than it has to be. I did it for the following reasons:

  • The sequencer has to be about 1/2 a euro rack wide just to accommodate the pots and switches (like 8 or 9 inches wide), so I didn't have to make it small.
  • There are so many flying wires and connections, I wanted some room to work, so I spread out all of the locations where the off boards connections were to be made.
  • I hate working and soldering in tight places.
JH, a, would have put it all on a "single 40cm x 30cm PCB" I'm sure, but I don't mind the extra cost for a little luxury when soldering :)