Something New from Doc Sketchy

[Dcramer]

Sure, but give me a week or so to sort out all of my other orders (plus I ran out of some parts so I'm waiting on a Digikey order before I can finish a couple of orders).

The 3-PCB stack is nice, and only a couple of inches deep, but it's also a couple of inches thick. The single PCB is much deeper but is only one board so not very thick. The only thing is that I can only connect wires to the single PCB with vertical pin headers, whereas I can use sideways (90-degree angle) pin headers on the board stack, so the connections do not add to the thickness of the stack. The vertical connections render the single PCB almost as thick as the stack. Plus, it's just a lot nicer for me to work with the smaller boards.

The Euro panel for this module only needs to be 4HP to accommodate all the stuff, but it should probably be 8 or even 10HP to accommodate the width of the board stack. I'll work up a couple of designs and you can tell me what you think. There is probably no virtue in making it too small, because it's harder to wiggle.

Sounds awesome Doc, 3 stacked boards would be great and there’s no issue with panel width for me, 8-12 hp is fine!

[Dr. Sketch-n-Etch]

I have 75mm to the back of my diy 3U case, how deep is the triple stack in total? 10Hp or even 14 Hp is cool with me, i like to have room to wiggle, defiantly an effect you want to wiggle! your prices are very reasonable considering the price of what other Frequency shifters go for, never thought i ever get my hands on one.

Concerning my prices, I do this for fun. I could never sell things at this price if this were my livelihood. The parts aren't that expensive, but this is all pretty time-consuming. I probably earn well less than minimum wage from all this stuff. I do this for gratification rather than profit. If I can make enough to cover my (ever increasing) Digikey and Small Bear costs, then I'm happy. Even so, though, most frequency shifters are a ripoff. The Encore one is a decent price, but many others are just ridiculous.

If you mount the stack perpendicular to the panel, then the boards themselves are 2.1" wide (about 53 mm). Only the top board has any panel connections. Depending on how you do the connections, you could almost mount the boards flush with the panel. However, you'd probably need at least 0.5 to 0.75 inches (10 to 20 mm) of clearance to bend wires. It can be done, but it's a bit of an engineering problem. The panel PCB can be fitted with 90-degree-angle pins. You could even put the pins on the component side of the panel PCB and bring the wires around the sides.

The board stack is probably a little bit less than 2" thick, but it has to be mounted to something. If the panel were 12 HP wide, then the boards could be mounted parallel to the panel. Also, the panel PCB could be organized with mounting holes that would line up with the corners of the Freak Shift PCBs That would probably be the way to go. The wiring may make the back of the unit a little wider than 12HP, but it would be deep enough that it probably wouldn't interfere with the modules on the sides. It might not like being mounted at the end of the rack, though, because the wires might impinge on the wall of the case.

If you want to do it this way, I'll work up a design. I think our friend J Daniel Cramer would also be interested in this design, since he wants a eurorack version (and he says that HP is not an issue).

[McIntosh]

It shows that your in it for the love and enjoyment it brings to people, not enough of this attitude, your a rare breed.

"If the panel were 12 HP wide, then the boards could be mounted parallel to the panel. Also, the panel PCB could be organized with mounting holes that would line up with the corners of the Freak Shift PCBs That would probably be the way to go"

This sounds great I'm in. I wait with baited breath.

[Dcramer]

Yup that sounds good to me!

[soup]

The board stack is probably a little bit less than 2" thick, but it has to be mounted to something. If the panel were 12 HP wide, then the boards could be mounted parallel to the panel. Also, the panel PCB could be organized with mounting holes that would line up with the corners of the Freak Shift PCBs That would probably be the way to go. The wiring may make the back of the unit a little wider than 12HP, but it would be deep enough that it probably wouldn't interfere with the modules on the sides. It might not like being mounted at the end of the rack, though, because the wires might impinge on the wall of the case.

First off wonderful work Doc. It got me dreaming of a building a frac version. Frac's a bit different than euro but I just don't see how 12hp would be possible. There's only a little more than 4" between the rails. Parallel to the panel I think you'd need at least 28hp.

[KSS]

Frac has 4-1/4" between the rails and at least 6-1/2" deep. Use of the available depth is how it would work for all classic Blacet -and PAIA- Frac boxes.

[soup]

I know the frac dimensions, it's euro I'm not too sure about. I have a number of shallower cases so was daydreaming about pcb's parallel to the panel and figured it would have to be 4 fracs wide. It was the 12hp number I was confused about, I think that's only possible if the pcb's are perpendicular to the panel and the module is at least 6" deep?

[Dr. Sketch-n-Etch]

I messed up.

The boards are 5.6" long, so they cannot be oriented vertically behind a eurorack module. They would have to stick straight back, and would therefore require about 6.5" of clearance behind the panel (to also accommodate the panel PCB and the mounting hardware).

Gawd, I hate eurorack.

Bottom line: Unless you have a super-deep case, forget about this for eurorack. I may have to get proper SMD boards made, and then we'd be in business. Maybe I'll pursue that with a local manufacturer, but it won't be until the spring.

Also, if you want to make an order, please send me a PM. In the PM, please give me your email address (so I can send a Paypal invoice) and mailing address (so I can estimate shipping charges). Thanks!

[McIntosh]

Bummer, I was so close!
yeah, Eurorack is a bit small but it's the format I choose, really wish Intellijel would jump on this, or whoever, I would be over the moon if you pursued a SMD version. I have no problem waiting, he who waits....
Appreciate your efforts on this!

[Dcramer]

It still works for me I’ve got like a foot back there!

[Dr. Sketch-n-Etch]

Hey Dcramer, drop me a PM with your email address and I'll set you up with a Paypal invoice.

[Dr. Sketch-n-Etch]

So, lastnight I had a dispiriting experience. I bought 100 of each value of capacitor I use in the Dome filters of Freak Shift and set about sorting them into the 41 bins (from 980 to 1020) for each value. I had reasonably expected to find, for 5% tolerance caps, a more or less Gaussian distribution of capacitance around a mean value of 1000. What I actually saw, at least for the 1nF and 10nF caps, were a ton of values less than 980 (from 950 to 980 -- they still fall within the 5% tolerance, but they are all clustered at the low end of capacitance). I don't have bin bags for any values less than 980, and my "<980" bags are now chock full of caps. That means that I'll have to create up to 30 more bags and sort those caps. This is becoming a little bit untenable. Probably what I'll do is combine caps into bags covering two adjacent values (say, 993+994), which will only give me 20 bin bags, and then I'll create 15 more bin bags for the low values down to 950.

It really doesn't matter too much. If I have a lot of caps at, say, 970, or 960, then as long as all the values are at that mantissa, I can still build very accurate Dome filters from these caps. I have an Excel model of the filters which allows me to put actual cap values in and look at the phase displacement error, and I find that even if the caps fall within a range of capacitance of up to +/-0.5%, I can still get very close to the minimum phase error. For example, if I have, say, four 10nF caps at 1002, 1004, 1005 and 1008, the model allows me to place those caps in specific locations within the filter circuit to minimize phase error (typically, that entails putting the 1002 and 1008 on one side, and the 1004 and 1005 on the other side). It just adds to the time required to stuff the filters.

It just sucks that the cap values are skewed low, as it just makes my job that much harder.

[devinw1]

Ah, yes that is fun stuff. I have seen this sort of thing, sorting extremely tight tolerance ceramic sleeves for a fiber optic project and with wavelengths of sorted laser diodes.. You expect a Gaussian distribution within the stated bounds, but then it turns out that the manufacturer is dipping into to the distribution to either sell another company large amounts of a certain range or something similar, which skews the distribution. There are other factors that can skew the distribution, but that is certainly common. It may very well be that if you buy a huge bag of them from a different vendor (digikey vs mouser or something) that you get a way better distribution, but you're gambling of course. And you don't want to be buying thousands of caps you'll never use. Ayiyi!

[Huba-Swift]

I recall when I was building Jurgens Freq shifter's dome filter, the capacitors were measured and two resistors were chosen for each one based on that value, eliminating the need for precision capacitors. Is this possible or is your use case different?