OrCAD

After many years of making do with either restricted commercial software (Eagle) or old tools long out of support (P-CAD) I am now firmly entrenched in the OrCAD camp.

Time to put mouse to keyboard and get some new designs out!

Done (Really)

Following some gentle prodding by a fellow AP enthusiast (thanks Andy!) I finally found inspiration and time to sort out the auto-detect capability of the AP control program S1.EXE.  It turns out I had missed off a couple of logic paths, and following careful reading of the S1 service manual (again) I found that all I needed was a simple one-line fix in the Verilog..

Yay!

-- Neil

Cables

Next task is to make up some XLR cables.  I found a few metres of Gotham GAC-3 in my junk store and so set about an evening of soldering up some test leads:

A quick test on the SystemOne showed them to be as good as the internal "GENMON" loopback test.

On a related note, by setting the analyser input filter to 22Hz-22kHz it reads down to 0.00034% THD+N.  Awesome!

--Neil

Done (Mostly)

So there is one wrinkle in my System One interface card: the auto-detect doesn't work for some reason.  I'm pretty sure the logic is all there, and I've tested it with QBasic test programs.  I can start the AP software (S1.EXE) from the command line and specify the interface card address with the /I option and then everything is fine.  So at this moment in time I don't fancy digging in deeper to fix this minor issue.  It doesn't stop me using this amazing machine.

A side note: after setting the SystemOne to internal loopback I measure the THD+N as 0.0014% on both channels.  That's a decent 8dB better than the 8903B (about 0.0035%).

--Neil

Analyze This!

It works!  The damn thing works!!

Well, okay, there were a few bugs in the Verilog, but checking out each block soon ironed out those.

Ta daaa.... Start up screen:

Eagle-eyed readers will note the flat grey cable poking out the case - that's the ISP download cable for the Lattice CPLD.

A quick graph showing that scans are working (needed the external interrupt line wired up properly - took an evening to fix that little bug):

Note that the performance at the moment (THD+N @ 0.0033%) is about on par with the 8903B it currently sits on top of:

But that's with a less-than-ideal coax cable - I can shift the TND+N over 0.01% just by handling the cable.  I think I really need to look into replacing those BNCs with proper Neutrik XLRs.

The PC that I am using is an Advantech industrial control PC in a small case.  It has a 486 processor running DOS 5.00.

For now, this is just what I need.  Perfick!

--Neil

Stuffing for Christmas

So, yesterday's post showed the bare PCB.  Today we have the stuffed board ready for bring-up:

During bring-up I'll be adding each IC in turn, making sure it is doing what it should be doing.  The last one with be the CPLD - you can see the in-circuit programming header top-left of the board (a couple of the pins are also used to set the board address).

I've also checked the board fits into the PC. The next version of the PCB will be gold-plated.

--Neil

New project just in time for Christmas

I've started work on the next project.  Well, "started" in this case means getting the first prototype PCB back from the fab:

Took a while to get everything to fit, and the release version will be a little bit simpler.  As to what it does....I'm sure there are some folks out there who will be able to work out what this is for!
--Neil

Ring....ring......

Latest project on the bench coming along nicely.  Prototype built and tested just in time for this year's synth-diy meeting in Cambridge.

 

Ring modulators are fun!

 

LEVIN is on the bench!

Finally, after some time to get things just right, the next CESYG module is almost ready for release.  LEVIN is an LED Voltage Indicator with a seemingly simple function: show an input voltage on a large LED bargraph.  I think it does it rather well.  Here's the first completed module after testing:

Note that in the above picture it looks like several LEDs are on at the same time, when in fact that is simply due to the LEDs sweeping upwards while the camera was capturing the image.  To keep the current consumption under control the display is always in dot mode.

All that remains to do is finish writing the manual, get it printed and bound, and then package up the first batch of modules ready for shipping.

--Neil

Slow news.....

They say no news is good news, right?  Well, progress at CESYG has been fast and slow.  First and foremost, the SuMIX project has been shelved - but the idea lives on and may resurface one day. On a more positive note, the BUS5 and BUS5X are now available to provide a 5-channel routing system within and between modular synth cases.  Finally, design work is progressing well on the next CESYG module.  Nothing more to say on it yet until I get a prototype working on the bench, and then I'll post piccies to tempt you!

--Neil

SuMIX Headphone Amplifier

Tonight I've built the headphone amp, and tweaked the design to achieve lower distortion (noise is now the  dominant factor).  30R headphones you're looking at around 0.01% distortion at 1kHz when pushing 120mW into the headphone (unlikely!). For 600R headphones you can go even louder with even lower THD, and then I'm starting to reach the limits of my HP distortion analyser (below about 0.005%).

--Neil

First panel mock up of SuMIX

As hinted at earlier, the next module I'm working on is an output summing mixer for the Eurorack format.  And here is the first view of the front panel.

Ignore the blobs on the corners, those are just for aligning the artwork with the panel mechanics.
Also this shows representations of the panel furniture (knobs, leds, jacks, switches) to give a better impression of how it might look.
And thanks to Cody for his useful input :)

DuaLFO Goes Even Slower!

After some discussion on muffwiggler I ran a test to see how slow the DuaLFO could really go with a negative CV input.  Internally it is limited to -0.6V, so I applied -3V just to make sure.

Using my Heath-Robinson-style low frequency counter I measured the period to be about 8,833 seconds, or about 2 hours 27 minutes.  Now that's slow!!!

On to the next one

Now that the DuaLFO is finished and shipping it's time to start work on the next module.  This time it's going to be an audio mixer.  And I think it's going to be very interesting.

DuaLFO Now Available

The first (short) run of DuaLFOs is complete and ready to ship.  Get'em while they're hot!

Auxiliary LFO Equally Slow

Last night I ran the same tests on the Auxiliary LFO.  It has the same maximum frequency of 46Hz, but if anything it is slightly slower than the Main LFO.  I recorded a run of 1 hour 35 minutes.

I think that's slow enough....

Main LFO Goes Slow

Last night I finally characterised the frequency range of the Main LFO of the DuaLFO.  And it is quite a range!  Highest frequency was 46Hz.  Pretty good.  The lowest frequency?  Well, probably better to think of the period rather than frequency.  In this case I measured the period to be 1 hour 25 minutes.  Or, in frequency terms, that's 196 micro-Hertz!

SSssssssllllllloooooowwwwwww...................

Next is the Auxiliary LFO, but I'm guessing that it will have similar performance.

DuaLFO #1 Comes Alive

At long last, and after a couple of minor PCB mods (ahem....) the first DuaLFO is completed, tested, and fully operational:

As you can see this first run of modules is using the Cliff KM15B knobs, and in practice they look and feel great.  The SYNC pos/neg edge button has a nice red Rean cap, and the bicolour LEDs sit nicely in the panel.  I think the graphics worked out well too - the MetalPhoto finish is very crisp.

This is an 18HP panel, offering comfortably spaced controls, jacks and indicators.  While I can understand the benefits of squeezing modules into tiny panels to get a lot of functionality into a small space, it actually takes a lot of skill to design a small panel well.  The best examples of this I know of are the Tektronix 7000-series scope and TM500-series test equipment plugin modules.

The last task to do before these modules are ready to ship is to finish the user manual and get it printed.  If that works out as well as the panels then I think its going to look quite good.

Knobs!

Design of the first CESYG module is progressing well.  The PCB is nearing completion, the order for the front panels has been sent off, and most of the parts are in.

One aspect that needs finalising is the choice of knob.  After much searching and comparing, I've whittled the choice down to two options.

Cliff KM15B

This has a very clean black-on-black look, with diamond-cut lines on the top of the knobs.  From the Cliff website:

Knurled aluminium shell on plastic inner. Standard finish is black anodised to a very high standard.

 Nice.

Sifam TP 150 006

 

With a matching red cap with a white dot - continuing the round theme of the panel graphics - this looks very red and demanding attention from the eye!  It is definitely bolder than the Cliff knobs, but is it too bold?

But.... which to choose?  Choices...choices....

Dual LFO Front Panel draft

Time for an update while I'm refining the PCB layout.  Here is the first draft of the front panel.  Its 18HP wide, and I confess I'm not a graphic designer!

The four input jacks are on the left.  Then the CV pots and the Sync mode switch.  Next column we have the Frequency controls, LEDs, and the manual Shape control,  Finally, on the right are the five output jacks.

I need to experiment with knobs and refine the layout, but I think its quite a clean layout.  The intention is to get this screen printed or perhaps Metalphoto'ed.