System design

As the PCB and panel layouts progress, it's becoming more clear I need to standardize some things for ease of assembly and my sanity. To that end, I've taken all the power regulation/conditioning off of the main module boards and will use a common power PCB which can be built to work with any of the modules. The sequencer is the exception. It will have everything on its two PCBs. This will allow for more compact boards.

In what little spare time I had today at the office, I was working on the PCBs in progress which are:

VCO
VCA
VCF
EGs
Sequencer

I'm busy trying to get these PCBs done so I can get the next batch of prototypes ordered.

Nyle got the Sequencer working.

This was great. I'm traveling and when I got off the plane, I saw I had some voice mail. I call in and I got this weird synthesizer sequence playing for about 30 seconds. I thought this is a weird prank call, and then it hit me. It was probably Nyle and he got the sequencer working. I hit "call caller" and sure enough, Nyle got the sequencer going already. There was one more schematic error, very embarrassing. I had the run stop switch pole to the wiper of the frequency pot. This then directly connected the wiper to the emitter of the UJT when the switch was supposed to connect to the charging capacitor. That was messing with the oscillator as well as shorting the pot wiper to ground through the UJT when it switched. That torched the pot and UJT, go figure... The image on top is how I had it. The image on bottom is how it should be.

Anyway, Nyle figured that out and fixed it, connected it to a VCO and played me a tune. It was probably the most fun voice message I ever got. So, he is going to verify the other circuitry then off to the prototype PCB shop.

One suggestion Nyle made was to rearrange the pad order on the switches. He thought it made more sense to put the "wiper" or "pole" of the switch in the middle. I agree. It's a simple fix for the rearrangement, but it means having to re-do some layout that use switches, like the sequencer, and EGs. The pots already have the wiper to the center pad.

I think I'm finally happy with the VCO layout. I want to add a couple of helps to the silk screen, then I  think I'll send it in.

Anyway, the resurrection is moving along!

Some progress.

Work and life have been taking some priority the last few days as well as two hard drive failures.

Nyle came by today and brought my O'scope back and took the prototype sequencer, VCO, and an EG. He's been great with this project. I think he's as excited as I am and in some ways, probably more. He's going to go over the sequencer and bread board the upgrades to make sure they work. So, I'm going to hold off on ordering the next generation sequencer til he's had a good look.

I've been tweaking the VCO layout for no really good reason. I've got an output buffer board designed and gave the schematic to Nyle for his approval. I used the same 741 OpAmps in keeping with the vintage feel. Since it is a daughter board and an optional addition, I don't think it detracts from the VCO itself. I've also adjusted the panel layout just a bit to make sure the mounting holes don't interfere.

I've got the layout for the VCF pretty much done. I just need to make sure the mounting holes work on the panel.

I've been working on the final layouts for the EGs, too. I'm still conflicted regarding how to handle the triggers and gates for the EGs, sequencer, and the other couple of modules which use/generate S-Triggers. Right now, I'm leaning for the daughter board option. The circuitry is very simple and I can put the convertors on the daughter boards and then give the builder the option. The other way is to have a separate module. For my system, I want the separate module. It keeps the other modules authentic and true to Nyle's original idea, but will let me interface with my other system when I want to.

I think I can get the VCOs and VCFs ordered this week. the EGs should be soon after. I'm going to wait on the panels until I get these next PCB versions finalized, then I'll order some VCO, VCF, and EG panels.

I'm still hoping to have PCBs, etc. up for sale before the end of the year, and if all goes well, maybe a couple of months for the first VCOs, VCFs, and EGs.

More prototypes ready to order

I got the sequencer board stuffed and found an error in my schematic. It was in the logic section and was messing with the sequencing. I was able to confirm the flip flop logic is fine. I was not able to fully confirm the internal clock, however, since the production version uses an upgraded voltage controlled clock, I'm not going to worry about it. I did manage to burn up two pots however. Whenever I turned the pot to full CW, the pot smoked. This is weird because when I had a trimmer soldered in, the trimmer didn't smoke. I think in all my unsoldering/soldering around the clock circuit, I must have shorted something which was causing the magic smoke to be released.

So, I've made the corrections to the next version sequencer, I've updated the VCO to be two boards now (see below for details), and the VCF to be two boards. This lets me mount the VCF board parallel to the 14 Hp wide panel which I think will be satisfactory for most euro rack and frac rack users.

The VCO will now be two PCBs with an option to add a third daughter board with output buffers and output "levelers". I'm doing this because I figure there will be those who want the VCO outputs to be +/- 5 volts. As it stands, the pulse is about +/- 7 volts, the sine +/- 6.5 volts, and the saw and triangle about +/- 2.5 volts. This way, I preserve the heart and true Steiner VCO on the two main boards and then can add a third which will allow it to interface better with modern stuff.

The VCO is one of the more complicated PCBs in the system and has a lot of components. I wanted to reduce the size of the PCB in case anyone wanted to mount it perpendicular to the panel or design a slightly narrower panel. I was struggling with this until I thought I could just make the power regulation on a separate board and cut an inch or two off the main board. Right now, the panel is 28 HP but the main VCO board is now 4.5 inches wide and 3.9 inches tall and would fit behind a 23 HP panel, saving about an inch.

I'm going to follow this approach for as many of the PCBs as it makes sense. I'm also going to use the daughter board idea for some of the trigger conversion circuits, too. I think that preserves the ressurection idea for the Steiner system, but also allows a user to integrate it into an existing system with less headache.

I wonder how many people will actually buy/build these modules... I guess even if it's only one, it's worth it to do it right.

VCO progress

With some great help and suggestions from Nyle, I think the VCO is good to go. I had to change a couple of resistors and, sorry, added two more trimmers. One will set the oscillator frequency range. You input 0.0 volts to the CV in and then adjust this to get the lowest frequency you want to have. Then, input 5 volts and make sure the high end is high enough. When I used about 0.1 Hz for the low end, the high end was inaudible to me.

 There is no high frequency trim and Nyle and I discussed this, and in the end, I don't think we will add one. those with perfect pitch might be annoyed, but then, no analog oscillator can be perfect. I have two or three VCOs with high frequency trimming and I found that if you trim the high end for good tracking, it can tweak the tracking in other parts. On a really quick tune job today, I got it tracking almost perfectly at about 440 to 880 Hz. At two or three octaves higher, I was off by about 1%. I was pretty pleased with that as from note to note, I really couldn't tell.

Nyle also wanted to add an additional possible trimmer to the triangle shaper. The 2N5163s they used were varying wildly and there were some which would not allow you to properly shape the triangle wave with the default trimmer. He found if you take the trimmer on the collector and put it on the emitter, you could trim the triangle wave. He wanted to make sure the end user could do that if the FET was problematic. So, I have trimmers on the collector and emitter. In normal building, the trimmer on the emitter will be jumpered to ground. See the partial schematic above. So if the FET doesn't let you properly shape the triangle, take out R63, jumper it to +12, take out the jumper to ground at R72, and install the trimmer. Or, you can just install both trimmers and start with R72 full CCW (shorted to ground). That said, the 2N5246s Nyle used in the later VCO cores haven't given me any grief and I've tried several random ones.

I tested the sync, and it works perfectly. I didn't try trimming the Sync Bias though. The only thing left really is the mismatch in output gains. I'm honestly tempted to just leave them. I've got enough trimmers already!

I've got the sequencer about half stuffed. Nyle found a couple of errors in my schematic capture, and thankfully I can easily take care of them when building the prototype. It may also explain why my simulation didn't work right, too...

Next up, I'll give the VCO a final check and blessing from Nyle and order some more boards. I want to build the VCA next and then get it's next revision and the classic Steiner filter boards off for prototype. I also need to get the type III EG and the next revision of the type I/II off.

I've checked the VCO panel and the VCO PCB will fit fine parallel to the panel, so shallow euro rack guys be happy if you want a Steiner VCO. Sorry, but it will be 28 HP wide. I could make it more narrow, but then you  will need about 6 inches clear behind it. The panel would also seriously depart from the original layout and I want it to be a close as possible.

Enough for now.

Expensive, but exciting.

More on the trimmers

I thought I'd better clarify a bit on the trim pots. The VCO proper has 10 trim pots. Here is the Prototype PCB layout:

Min pulse width
Max pulse width
Phase sync trim
DC offset for the sine output
Sine shape
Sine shape bias
Triangle shape
Volt/octave trim
DC offset for the saw output
DC offset for the triangle output

More than other VCOs I've ever seen or built, but reasonably straight forward to set. The single turn trimmers visible in the photo on the edge of the PCB are simply surrogates for the panel pots. I used these because I thought it would simply be easier to deal with than a bunch of big pots with flying wires.

The next step for the VCO is to verify I can actually set the V/Oct trimmer and get it to track. I would like to verify the sync input, too.

Before I post the schematics, I need to add a provision for an onboard LED to provide a minimum load for the regulators if needed. Then as soon as I verify and triple check the component values, I'll order what I hope to be the last prototype and then this one should be ready to go.

There are only two more modules which are more complicated. The other ones, even the filter, are really pretty straight forward and I think they will turn around fast. The sequencer and the frequency divider will take more time. The sequencer just because of quantity, the divider because of complexity.

I think there is real reason to get the sequencer done soon if not next on the list. Nyle is very interested in having the sequencer well documented and he is excited to get it done, more than enough reason for me to put some effort into it. After that, I think the EGs, VCA, and VCF definitely need to be done as they are the core of any modular. The others are pretty simple and will come later.

It's Alive!

Thank you Nyle Steiner! (The top blurry cell phone photo is Nyle in front of his workbench.) I stuffed the newest VCO prototype and it basically worked first time. I was using a wrong value for a trim pot which prevented the triangle wave output buffer opamp from working, but I got that figured out and it works great! Here are some more crappy cellphone photos of the screen of a DSO Nano O'scope (long story but I loaned my bench scope to Nyle for a couple of weeks). The photos don't do justice to the screen of the Nano. Very cool little scope. I think this is the first new Steiner Synthesizer Company VCO in 30 years give or take. The pulse and sine outputs are higher than the triangle and saw outputs, so I'm going to triple check all the component values to make sure I'm doing this right. I'm also a little lost regarding the adjustment of the min/max width trim pots for the pulse wave. It's another question to ask Nyle.

Now, on to the sequencer prototype. I'm way excited again about this whole thing.

I've also got the rack/power distribution system figured out. I got the new power plugs and they fit perfectly in the existing cutout. I drilled the holes for the mounting screws and they look and work great. I don't know if anyone but Nyle and I will use the rack system, but it sure looks nice and I think will work as good as they look.

Another odd note. I decided to use adjustable regulators for the power section so I could get an accurate voltage on each board. I discovered through some kind help on the SDIY list I have to have a minimum current draw to effect regulation. I'm going to have to add some small load on each regulator because some of the circuits may not meet the worst case current draw to establish regulation, particularly on the -10 volt side. So another task before the next round of PCBs is ordered. Hopefully though, the next batch will be the first production boards.