Euro rack Moog 914 clone DONE. PCBs ordered.

[dingebre]

Update 06/08/2020

Just a note to suggest you have a look at the Moog 914 FFB pages on my site:

http://analoguerealities.com/projects/f ... -bank-914/

I re-visited the project, long story, and I suggest some simple corrections to the build. Some apply to the inductor version, some to the GIC version, and some to both. There are really no significant changes, just some capacitor values in the LP (inductor) and HP (both) and a couple of resistors and capacitor changes on the I/O PCB. Simple as they are, they make a big difference in the behavior and response. Much closer to the original Moog.

Anyway, have a look.

David





Update 12/17/2014
My new web store is the best place to buy the PCBs. Start at:

www.analoguerealities.com

click on the Catalog and SDIY "key", navigate the menu to the Moog modules, 914 FFB. It's the only one right now, so should be easy. The checkout is through PayPal.

David



************************************


Good news, the GIC and Inductor versions are both finally working. I ordered PCBs today. If anyone is interested in buying a set, email or PM.

GIC - $70.00 per set (A little expensive, but they are 4 layer PCBs to minimize noise and hum)

Inductor - $50.00 per set (Will work with either the large or small core 5 Henry inductors Carsten has made)

A couple of quick comments on construction. I'll put more detail together if people buy the PCBs.

A set consists of 4 PCBs, a top, middle, and bottom PCB adn an I/O PCB which mounts to the bottom filter cell PCB. They have PCB mounted pots, Panasonic EVU-E2JFK4D54, DigiKey P3U0503-ND. You will most likely need to trim the shaft a little depending on the knob you use. the HP and LP pots are not PCB mounted and can be any smallish panel mount pot. I've calculated the values for the capacitors which are on the silkscreen. There are two choices, one for an "exact" cutoff frequency, one that pushes the cutoff either up or down about 6% to mimic what it appears Moog did (as well as Jurgen Haible on his version).

The construction is not bad. The calibration is the problem, here. Ideally, you want the overall "Q" to be about 3.7 (per Yves Usson's testing of an original 914). The trimmers will adjust that, but it's not so straight forward to tell. I've just set my trimmers to point to about 10:00 or 11:00 and the result is very pleasing. you don't want too high a "Q" or you lose some of the harmonics between the cells. Too low and you miss the nice resonance when the input attenuator is fully CW.

Be warned, both versions are kind of pricey to build. For the inductor version, you need a set of Carsten's inductors. for the GIC, you need 29 of the best, lowest noise, op amps you are willing to buy. I ended up using the OPA2227, about 4.50 each for 10-99 pieces. Probably way overkill. Jurgen used an MC4558 at about $0.40 depending on the exact one you pick.

So, enough for now. Please feel free to contact me if you are interested. I've got about 2 weeks til the PCBs are in. In the mean time, I'll work on a BOM and some construction notes.

END EDIT
################################################


I've been working on a Moog 914 fixed filter bank clone using GIC (generalized impedance converters) and inductors (Carsten Tonsmann at www.Analog-Monster.de custom wound)

I combined ideas from Yves Usson, Jurgen Haible, and my own doing to put together what I think is a faithful recreation of the original.

Good and Bad:

Good:
Euro/Frac rack compatible
PCB mounted pots (except for two)
Can run on +/-12 or +/-15 volts
Doepfer and MOTM power connectors
Two versions:
1. GIC version using op-amps to simulate the impedance of an actual inductor. All the good from a "real" inductor and none of the bad, like interaction with stray electromagnetic fields.
2. Inductor version use real inductors with the nonlinearities and other "features" you get with a real inductor.

Bad:
It uses 4 PCBs and they mount perpendicular to the panel, so you need a deep cabinet.
The inductor version is very sensitive to stray electromagnetic fields. I can't find a place in my cabinet far enough away from the power transformers to eliminate a very high level hum. But, if you use a toroidal transformer or if the transformer is a couple of feet away, you should be fine.
The GIC version uses 4 layer PCBs so they are more expensive than the standard 2 layer, double sided PCBs.

Idea:
I wanted my own version of the 914. I built Yves Usson's excellent version, but it is totally an active filter circuit. It also uses a PCB much more suited for a 5U module, not euro or frac. I bought circuit boards for Jurgen Haible's version, but again they are more suited for a 5U format. With all due respect, Jurgen also missed an error on the Norlin schematic where a resistor in parallel with each second band pass cell was omitted. Without this resistor, the "Q" in each cell is too high and you get too much resonance in the filter.

So, I borrowed Yves' (and Kazike at COTK although he brings every cell out) idea to also bring out every other cell as an "even" or "odd" cell. I also decided that a back to back GIC low pass cell was not ideal. So, I designed a salen-key low pass filter cell similar to Yves' and Jurgen's design but with standard component values so no need to put resistors in series or capacitors in parallel.

I use the original Moog input and output amplifiers for the combined filter outputs, but do a simpler op-amp (either a single stage or dual stage if you want) output for the even and odd outputs.

There is an interesting feature in the 914 (Jurgen discusses this, too) output amplifier, a high pass filter with a cutoff at about 70 Hz at the input to the output amplifier. I included this filter as did Jurgen at the input to my even and odd op-amp output stage. It's already in the Moog stage. I think Moog included this to help filter out stray 60 cycle and to eliminate the muddy sounding very low frequencies. This filter is intended to help form certain instrument sounds by emphasizing certain resonances, so I think they wanted to get rid of "boomy" bass when using the filter. But what do I know. I'm looking at a 30-40 year old circuit trying to second guess what the design engineer's intent was. But this is my story and I'm sticking to it.

Update:
OK, after many, many hours and lots of prototype PCBs, I think I have a good handle on the filter. I just ordered what I hope is the final revision for PCBs for the GIC version. I went to some very extreme measures to minimize outside noise and hum pickup, like a 4 layer board, ulta low noise op-amps (OP2227PA), etc. I think I can safely say any noise is due to the original filter design. I think the inter PCB connections carrying the summed output signals may derive benefit from shielded wires due to the low levels from the resulting summed filter cell outputs.

For comparison, I did some quick "ear" checks against my Yves Usson 914 filter clone and I was very happy. I'll work on some more analytical comparisons over the next week or so.

I need to think about the Inductor based version a bit more, but I think I have the final version of those PCBs ready to go, too. I will buy a set of last prototypes soon and will post here as soon as I have them built.

David

[Isaiah]

[Isaiah]

I've been following David's progress here (just in case anyone missed it):

Http://electro-music.com/forum/topic-58676.html

[Isaiah]

I've been following David's progress here (just in case anyone missed it):

Http://electro-music.com/forum/topic-58676.html

[felixer]

I need to think about the Inductor based version a bit more

to minimise noise you may want to put them in a mu-metal box. yep ... expensive.

[metasonix]

I've been working on a Moog 914 fixed filter bank clone using GIC (generalized impedance converters) and inductors (Carsten Tonsmann at www.Analog-Monster.de custom wound)

You, sir, are insane. Congratulations!

to minimise noise you may want to put them in a mu-metal box. yep ... expensive.

Nope. Mu-metal is absurdly overpriced and only suitable for special applications where weight and bulk have to be kept to a minimum. A cheap mild-steel shield works just as well, if it's made about twice as thick as the mu-metal shield needed.

[dingebre]

I've been working on a Moog 914 fixed filter bank clone using GIC (generalized impedance converters) and inductors (Carsten Tonsmann at www.Analog-Monster.de custom wound)

You, sir, are insane. Congratulations!

to minimise noise you may want to put them in a mu-metal box. yep ... expensive.

Nope. Mu-metal is absurdly overpriced and only suitable for special applications where weight and bulk have to be kept to a minimum. A cheap mild-steel shield works just as well, if it's made about twice as thick as the mu-metal shield needed.

To Metasonix: Many thanks, from one insane person to another, I take that as a high compliment

To Felixer: I think that for most applications, you can locate the module far enough away from a power supply to avoid the hum. The way I've done my power distribution just makes that basically impossible. I don't think the inductor version of the module is any more sensitive to stray 60 Hz than a reverb tank, so if you can locate a reverb tank in your system, the inductor version should work. I've put my prototype in a different rack and it seemed to work fine. Also consider the Moog version was not shielded, but I think Moog used a toroidal coil which is less sensitive to stray fields. Carsten has told me he hasn't had any problems in his system with hum which is encouraging.

As for actual "noise" I think there are issues with the input attenuators and large resistors in the audio path coupled with a final gain stage of about 230 introduces low level noise that is simply inherent in the system. Jurgen mentioned this in his write up and I think I agree that there will simply be some low level noise due to the design. But isn't that part of the "charm"?

I appreciate the comments.

David

[felixer]

I don't think the inductor version of the module is any more sensitive to stray 60 Hz than a reverb tank, so if you can locate a reverb tank in your system, the inductor version should work.

well, single coil strat-type pickups are still being used in guitars. and are a pita these days. in the 50/60ies there was much less electromagnetic dirt around. now, 1 guy carring some bloody 'smart'phone can ruin a session ...
weren't the original moog mudules shielded?
yeah, thick iron/steel does it too. mu-metal cans i see mainly in mics and preamps ... that's partly why all this old broadcast stuff weighs a ton ...
but if you go for a 'recreation', all that's prob pretty irrelevant. maybe the user decides to go old-style tape/noNR anyway
i must say that inductor based eq often just sounds ... nice. how big is the difference in character, all else being equal? because usually the designs are so different it's hard to pinpoint ...

[dingebre]

I don't think the inductor version of the module is any more sensitive to stray 60 Hz than a reverb tank, so if you can locate a reverb tank in your system, the inductor version should work.

well, single coil strat-type pickups are still being used in guitars. and are a pita these days. in the 50/60ies there was much less electromagnetic dirt around. now, 1 guy carring some bloody 'smart'phone can ruin a session ...
weren't the original moog mudules shielded?
yeah, thick iron/steel does it too. mu-metal cans i see mainly in mics and preamps ... that's partly why all this old broadcast stuff weighs a ton ...
but if you go for a 'recreation', all that's prob pretty irrelevant. maybe the user decides to go old-style tape/noNR anyway
i must say that inductor based eq often just sounds ... nice. how big is the difference in character, all else being equal? because usually the designs are so different it's hard to pinpoint ...

I don't think there were shielded. I've seen some detailed photos of the 914 and the PCBs are in metal frames, but not fully enclosed as far as I can tell. Of course, the photographer may have removed any shielding that was present for the sake of the photo.

[dingebre]

I stuffed the latest prototypes using Carsten's inductors and I re-did my whole power distribution. There is now no hum whatsoever and Carsten's inductors are great. I discovered one last "fitment" issue on the inductor PCBs and will order what I hope is the last set of prototypes tomorrow.

I also re-did the layout around the summing nodes and I think that has helped as well.

I will stuff the GIC prototypes tomorrow. I'm really quite pleased.

David

[dingebre]

Done...

Here are photos of the GIC version. I should have the final prototype inductor version tomorrow and I'll stuff as soon as I can. Work is ramping up and I'll be pressed for time. I think the low pass is a little hot, but that can be adjusted by making the summing resistor bigger. I'll try to get audio and a spectrum analysis posted as soon as I can make time.

[dingebre]

Good news, the GIC and Inductor versions are both finally working. I ordered PCBs today. If anyone is interested in buying a set, email or PM.

GIC - $70.00 per set (A little expensive, but they are 4 layer PCBs to minimize noise and hum)

Inductor - $50.00 per set (Will work with either the large or small core 5 Henry inductors Carsten has made)

A couple of quick comments on construction. I'll put more detail together if people buy the PCBs.

A set consists of 4 PCBs, a top, middle, and bottom PCB adn an I/O PCB which mounts to the bottom filter cell PCB. They have PCB mounted pots, Panasonic EVU-E2JFK4D54, DigiKey P3U0503-ND. You will most likely need to trim the shaft a little depending on the knob you use. the HP and LP pots are not PCB mounted and can be any smallish panel mount pot. I've calculated the values for the capacitors which are on the silkscreen. There are two choices, one for an "exact" cutoff frequency, one that pushes the cutoff either up or down about 6% to mimic what it appears Moog did (as well as Jurgen Haible on his version).

The construction is not bad. The calibration is the problem, here. Ideally, you want the overall "Q" to be about 3.7 (per Yves Usson's testing of an original 914). The trimmers will adjust that, but it's not so straight forward to tell. I've just set my trimmers to point to about 10:00 or 11:00 and the result is very pleasing. you don't want too high a "Q" or you lose some of the harmonics between the cells. Too low and you miss the nice resonance when the input attenuator is fully CW.

Be warned, both versions are kind of pricey to build. For the inductor version, you need a set of Carsten's inductors. for the GIC, you need 29 of the best, lowest noise, op amps you are willing to buy. I ended up using the OPA2227, about 4.50 each for 10-99 pieces. Probably way overkill. Jurgen used an MC4558 at about $0.40 depending on the exact one you pick.

So, enough for now. Please feel free to contact me if you are interested. I've got about 2 weeks til the PCBs are in. In the mean time, I'll work on a BOM and some construction notes.

[sonicwarrior]

I'll try to get audio and a spectrum analysis posted as soon as I can make time.

An audio comparison between the two versions would be very nice.

The inductors can be had from here? http://www.analog-monster.de/mmt914_en.html
Because I see no mention of an inductor offer there.
Only a very vague note in German in the 'shop' section: http://www.analog-monster.de/shop.html (Short translation: For customized inductors or inductor sets write an email to ...).

[dingebre]

I'll try to get audio and a spectrum analysis posted as soon as I can make time.

An audio comparison between the two versions would be very nice.

The inductors can be had from here? http://www.analog-monster.de/mmt914_en.html
Because I see no mention of an inductor offer there.
Only a very vague note in German in the 'shop' section: http://www.analog-monster.de/shop.html (Short translation: For customized inductors or inductor sets write an email to ...).

I think I can get the audio comparison done between mine and Yves' version.

Yes, you just email Carsten from his website and he'll get you a quote. He has a link to his email at the bottom of his home page:
http://www.analog-monster.de/index_en.html

And on the "shop" page:
http://www.analog-monster.de/shop_en.html

David

[Hirsbro]

why arnt there more fixed filterbanks in euro? I only know of two ADDAC and dopfer

[Isaiah]

dingebre
Good work, David!
How deep will an assembled inductor-driven module be?

[dingebre]

PCBs have arrived for both GIC and inductor versions. PM or email if interested. See first post for details.

David

[Dogma]

dingebre
Good work, David!
How deep will an assembled inductor-driven module be?

This, the ADDAC and i think EMW make one. Def under represented. Im really keen on doing this but I cant do anything without a mouser cart. If there is one Im in. If there is one could someone please PM me

[dingebre]

dingebre
Good work, David!
How deep will an assembled inductor-driven module be?

4.5 inches to 4.875 depending on how you mount the I/O PCB.

David

[dingebre]

Here are some quick specs on the Module:

Current:
Inductor: +20.5 ma, -18.7 ma
GIC: +210 ma, -210 ma (can change depending on the OpAmp used)

Depth: 4.5 in typical 4.875 in at most depending on mounting configuration of I/O PCB
Panel width: 28U, 5.53 in (141.9 mm)

David

[Cablebasher]

great news.

been after a eurorack diy fixed filter bank for ages.

a few questions that i thought i would post here rather than pm as I imagine others might be interested to.

Here goes:

- is a eurorack panel likely to be produced?
- i have never worked with inductors before, are they difficult to solder (are there any special tools needed)?
- is there any smt?
- which type is easiest to build (inductor or GIC)
- which type sounds closer to the orignal?
- what price are the customised inductors likely to be?

thanks and if these have been answered before and i misssed it.

[dingebre]

great news.

been after a eurorack diy fixed filter bank for ages.

a few questions that i thought i would post here rather than pm as I imagine others might be interested to.

Here goes:

- is a eurorack panel likely to be produced?
- i have never worked with inductors before, are they difficult to solder (are there any special tools needed)?
- is there any smt?
- which type is easiest to build (inductor or GIC)
- which type sounds closer to the orignal?
- what price are the customised inductors likely to be?

thanks and if these have been answered before and i misssed it.

These are great questions.

1. I have an FPD file (see attached) which you are welcome to modify to suit your style. I kept the font I've used for the Steiner modules I built and the general layout scheme. You should keep the overall pot and jack layout, though otherwise I can't say if everything will fit.

2. The inductors are not that hard to deal with. The "cores" are a bit brittle and you need to be careful when you mount them. I recommend using a plastic bolt and nut, but even then I broke a core. If you do, it's not the end of the world. You can buy replacement cores from Mouser.

One other issue is the wires are extremely fine and the insulation is some kind of varnish, so you can't really "strip" the wires. Carsten recommends holding the soldering iron against the wires a little longer when soldering the connection which will melt the varnish. I had trouble getting a consistent, reliable connection, though.

I ended up by putting the wire through the hole and then about a half inch from the PCB, I heated the wire and feed some solder onto the wire at the same time and then cleaned off about an inch or so of the varnish. I pull the wire back through until the clean portion is in the hole and I get a great connection every time. this also leaves some slack in the wire. No special tools needed.

3. No SMT. There are some decoupling caps that need to be small. You can put them on either side of the PCB, just do it before you mount the ICs or sockets. If you don't use a socket, you will need to put them on the solder side before the ICs and trim the excess leads really close. I installed the sockets, then pushed the cap between the socket rails, but you can do it which ever way works best for you.

4. Both are pretty easy. Minimal flying wires, two pots and the jacks is all. The GIC version has more soldering because of all the ICs. The big issue is calibration. I don't have a specific procedure. In the end, I just set the trimmers to about 10 o'clock which should be close to the "ideal" value for setting the "Q" of the filter cell.

5. Don't know how to answer this because I don't know what the original sounded like. I've compared both to Yves Usson's excellent design and I can't tell a difference. I'll be posting some audio comparisons hopefully later today with the inductor version and Yves' module. I think both versions are accurate reproductions of the original. The frequency response curves I've looked at are comparable to the ones Yves posted on his site and the one for the original on the Moog Archive web site.

6. Last batch I bought was about $310.00 for a full set plus shipping, but that was a year ago. I think it's very reasonable considering they are essentially hand wound. However, I can't speak for Carsten. Contact him through his website:
http://www.analog-monster.de/index_en.html
and he will give you a current quote.

Please feel free to post questions here or through PM or email.

David

[Cablebasher]

Thanks for the quick response. Should this panel layout work for either version (GIC or inductor)?

I am thinking the inductors might be the more fun build for me personally as it's not something I've dealt with before. Do you have any photos of a completed inductor board so I can get a feel for how much work is involved?

I am aware of yusynths and haibles modules but I know you have done much work on updating them and I am intrigued to see how the inductors are mounted.

Look forward to the sound samples and BOM.

Keep it up.

[dingebre]

One other note, on closer examination, I did not receive the I/O PCBs in the order. I'm looking into it and will get them either way, but it will be a couple of weeks before I have them. Sorry.

David

[dingebre]

Thanks for the quick response. Should this panel layout work for either version (GIC or inductor)?

I am thinking the inductors might be the more fun build for me personally as it's not something I've dealt with before. Do you have any photos of a completed inductor board so I can get a feel for how much work is involved?

I am aware of yusynths and haibles modules but I know you have done much work on updating them and I am intrigued to see how the inductors are mounted.

Look forward to the sound samples and BOM.

Keep it up.

The panel will work for both.

See attached for the inductor version