[BUILD] L-1 Vocoder build thread

[consumed]

all documentation at:
http://l-1.su/Vocoder.html


PROJECT NOTES

- five eurocard size (160x100mm) circuit boards (4 filterbank, 1 control)

- works on 15V or 12V

• +V 900mA
  -V 430mA

- no rare parts - THAT ICs and average resistors and caps

- expensive ICs:

• (25) THAT4301
  (12) THAT300
  (1) THAT1510
  (1) THAT340

module depth will be 80mm without or 84mm with backplate

Pads exist with 2.54mm pitch for CV ins and outs if you desire to add these to your panel.
Trimpots exist for each band, but they have very limited range.
If you want level pots for bands, maybe better to install attenuators on the carrier inputs.

(Original development thread: viewtopic.php?t=56280)

[consumed]

=====================================================================

Mouser Filterboard Projects (FOUR filterboards)

C1 caps: https://www.mouser.com/ProjectManager/P ... b43cafd04d
C2 caps: https://www.mouser.com/ProjectManager/P ... 417c337457
most of the rest: https://www.mouser.com/ProjectManager/P ... 8a1cf832db

*note! these mouser projects are currently missing:

• THATcorp ICs (group buy thread)
  2k tempcos (i substituted normal 2k 1% metal film resistors--use smallbear tempcos below if desired and remove 2K from the mouser project)
  samtec connector (samtec.com)
  820p, 680p, 560p C1 caps! (smallbear)

*other notes:

• all C1 and C2 caps are 5% tolerance and 5mm lead spacing
  all C1 10n and up are Panasonic ECQ-V film
  all C1 < 10n and >820p are polyester film
  all C1 - 820p, 680p, 560p order at smallbear (film caps, below)
  all C2 10n and up are Panasonic ECQ-V film
  all C2 6n8 and below are polyester film
  i used regular electros instead of tantalums for 2.2uF
  i increased quantities when price breaks brought the cost down (e.g. 200 resistors instead of 48)

=====================================================================

Mouser Controlboard Project

https://www.mouser.com/ProjectManager/P ... 7af334cdc4

*note! this mouser project is currently missing:

• THATcorp ICs (group buy thread)
  pots (smallbear)
  knobs (smallbear)
  samtec connector (samtec.com)

=====================================================================

Smallbear Shopping List




2K tempco / thermistors
NOTE: remove the 2K resistors from the Mouser project above if you order these!



=====================================================================

Samtec.com Shopping List
try to order these as free samples. otherwise, the shipping charges from samtec very expensive ($15 from US to US). newark also carries this part.



=====================================================================

please have a look at these and provide any feedback or corrections.



edit: corrected smallbear film cap quantities - mar 23

[Funky40]

much thanks for the thread.

A rough look at Farnell turned out that i would have to pay 500CHF/$ for the That ICs alone, when counting with 100+price on the 4310.
farnell prices taken from other country might be a bit better than CH.
couldn´t find that ICs at mouser


well, i´m a bit shocked
finally there´s allways a risk that a project isn´t running and one isn´t that good in debugging
This project calls for much and careful info upfront and possibly for Guys who mighta debug ones oopsy-work
so the question is also how much $ could a debugging job be ?.......at least for me thats now a question
ok, it´s gonna be a 850-950$ build or so am i wrong ? then correct me please.
while i just have my second Doepfer vocoder coming, i´m so much dreaming to have a real good DIY Vocoder
really need to get very precise infos and pictures on the modification with all the patchpoints. Thats what makes a vocoder funny IMHO

[consumed]

nate/samuraipizzacat29 is organizing a thatcorp group buy. you should pm him.

viewtopic.php?t=75755

[magman]

THAT IC's are available from Mouser, but for some unknown reason aren't found by the Mouser search engine.

For example, the THAT4301 can be found at this link

http://uk.mouser.com/ProductDetail/THAT ... vJiNOQhHcZ

Mouser works out at about 60% of the Farnell costs.

I suspect that the groupbuy will be even better. If we get 40 participants, then we can possibly get the THAT 1000 quantity rate, which is just $4.46, plus whatever handling and groupbuy charges that need to be added.

I expect I will be buying PCB set for my project stack, but Vocoders are rarely inexpensive projects.

Regards

Magman

[Luka]

much thanks for the thread.

A rough look at Farnell turned out that i would have to pay 500CHF/$ for the That ICs alone, when counting with 100+price on the 4310.
farnell prices taken from other country might be a bit better than CH.
couldn´t find that ICs at mouser


well, i´m a bit shocked
finally there´s allways a risk that a project isn´t running and one isn´t that good in debugging
This project calls for much and careful info upfront and possibly for Guys who mighta debug ones oopsy-work
so the question is also how much $ could a debugging job be ?.......at least for me thats now a question
ok, it´s gonna be a 850-950$ build or so am i wrong ? then correct me please.
while i just have my second Doepfer vocoder coming, i´m so much dreaming to have a real good DIY Vocoder
really need to get very precise infos and pictures on the modification with all the patchpoints. Thats what makes a vocoder funny IMHO

well at least with filterbanks / vocoders there are a lot of duplicated sections which means while the project looks really big there are only a few circuit blocks to understand

this project was always going to cost a lot of money - i was curious why people were frustrated with the pcb cost considering the parts will amount to a much higher price

[fluxmonkey]

to find THAT ICs at Mouser, select ""View all suppliers" from the Suppliers tab, select THAT from the list, and then search w/in That components. or those components, if you were an English major.

looks like the set of THAT ICs would run about $250. i don't mind paying for quality parts for a design like this; can only imagine (and never afford) what a simlarly loaded commercial product would run to....

thanx as always, Roman.

anybody thinking Buchalish panel???

b

[samuraipizzacat29]

^ parts + LABOR =

2 days of 16 hours working for moderate speed assembly

[PF]

anybody thinking Buchalish panel???

b

[Funky40]

i don't mind paying for quality parts for a design like this; can only imagine (and never afford) what a simlarly loaded commercial product would run to....

yup, my point is that whatever i build, i do not build it with a guarantee that it will work. but thats me

well at least with filterbanks / vocoders there are a lot of duplicated sections which means while the project looks really big there are only a few circuit blocks to understand

yeah, you´re right. ...gotcha

Mouser works out at about 60% of the Farnell costs.

when the groupbuy comes to 500 it would be more like 30% than compared to my swiss price. GREAT !
makes it much more appealing again

Much Thanks Guys !!
(patchable) Vocoders within modulars is one of the greatest patchexperiences for me. Just to say

[Luka]

i think we need to work out a good way to patch this thing out. it might take some reverse engineering. i dont use euro format so it is going to take some customisation to get this thing working anyway

[roman_f]

anybody thinking Buchalish panel???

b

i was, but 84mm depth...

[diablojoy]

i think we need to work out a good way to patch this thing out. it might take some reverse engineering. i dont use euro format so it is going to take some customisation to get this thing working anyway

+1 here
I am in a similar position.
I seem to recall L1 stated he left pads for some patching points, i can live with wiring pads but hopefully they are at 2.54mm spacing for headers .

[L-1]

Some corrections are needed. If you use parts not higher than 10mm you can achieve 12mm between boards. 1mm of empty space and 1mm for solder joints are needed. In this case module deep will be 80mm without or 84mm with backplate.

Gain trimpots have limited range (something like as in quad vca). I didn't trimmed them on mine unit, just made them identical. But I left this trimming possibility to allow some individual trimming for those who want.
Second trimming is distortion trimming in compressor, same as described in Microcompressor docs.

So if you want level pots for bands, maybe better to install attenuators on carrier inputs.

And if to make 20mm between boards, another between boards connectors are needed with longer legs. Samtec has them but can't say exactly which.

So if to make patch points and level pots, some additional panel is needed. I will not provide it. But can make design or provide design elements to be in one style.

[Funky40]

here is the info on the In/Out patchpoints. looks good. Thanks for pics LI
viewtopic.php?t=56280&start=300


@L1:
more build questions:
how are the boards connected to the control board ?
I just see the interconnection between filterboards.
My Q is leading to: can we attach each filter board separatedly to the frontpanel, each after the other,
to get a functiontest donne for each board ?
can the frontboard be tested alone ?
some infos in this regards would be very cool.

[L-1]

Yes, a control board can work alone or with any quantity of filterbank boards.
Interconnection between all boards via 6-pin pads 2.54mm spacing. I've found Samtec headers with long legs and used them. Samtec number ESQ-106-13-G-S

[samuraipizzacat29]

viewtopic.php?p=1043566

[consumed]

Some corrections are needed.

thanks aleksei, ive edited the first post.

[raisinbag]

here is the info on the In/Out patchpoints. looks good. Thanks for pics LI
viewtopic.php?t=56280&start=300

Cool I missed this. So we could basically run to a switched jack on a companion panel? That rocks!!!! Now I need a picture of Alex's and Muffs face for the second panel.

Ok so what kind of voltage would we run into those links an how would affect each band? (-5-+5v, 0-5v, 0-10v?). Maybe I am not understanding how to run cv to it so don't spit on me if I'm all outs whack.

And if we want to control the level of each cv on the separate panel, we could run pots in series with each cv in?

[Funky40]

And if we want to control the level of each cv on the separate panel, we could run pots in series with each cv in?

attenuators AND offset would be the choice IMHO,
so separate modules as A129-3 is would be a good thing.
have a look at the doepfer vocoder ( A129)

the buchlidian CV thingy PCB, i guess from negativspace( ? ), should be the right thing for such addon module.
in fact i have 15 soldered, but they have not sent me the the pots yet
i gonna do a DIY addon module for my doepfer vocoder with those PCBs.

thanks for Info alexsei !

[consumed]

ive been going through the BOM creating a mouser project, but its slow going because i cant find one type of 'quality' cap to cover all of the C1 and C2 values. (no go with polypropylene, mlcc c0g most but not all)

edit: aleksei, will you provide a schematic for this project?

[L-1]

I will make Farnell links soon. But caps I've ordered from Smallbear and Banzai. Panasonics.
Green chinese caps I used on the prototype. Work well but they are a bit higher then needed but anyway can fit if to tilt a bit. I've bought them as values pack on ebay. This pack covers all needs, 10 caps each value. Costs very cheap.
Something like that but check values before ordering. And values higher than 47n will be too big.
http://www.ebay.com/itm/470pF-470nF-280 ... 1c2c2221bf

[L-1]

edit: aleksei, will you provide a schematic for this project?

Yes will provide but a bit later - together with all docs.

[mechie]

Re patchpoints:-

I can see how the external control of this is tempting BUT

Each of the 24 bands has:-
An analysis filter and envelope follower giving a voltage out that represents the amplitude of that frequency band at that instant in time.
A reconstruction filter, allowing an identical band through (if any harmonics exist in this band!); this feeds into a VCA. This is normally controlled directly by that band's envelope follower so that a similar amount of carrier is allowed through to that seen by the analysis filter.

OK, so for patching, I see only a few tricks that are likely to be tried:-

1/ Scrambling - cross-connect the analysis channels to different reconstruction channels, giving weird results.

2/ Slew-limiting - making the reconstruction filters slow to respond, I guess this might sound a bit 'drunken' or blurred?

3/ Leaving the patching as designed and tee-ing in to all (or a selection) of patchpoints to play with unvoiced detector ideas - a seperate circuit that could, on detecting an unvoiced sound, would substitute the vocoder's output for a filtered noise source.

4/ (Depending on the smoothness of the overall filterbank profile) Using the reconstruction filters as a kind of 'String Filter' similar to the JH String Filter project, requiring all channels to be open all the time - maybe easiest just putting the same waveform in to both vocoder inputs and, possibly, killing every other channel to give a more 'peaky' response

As far as external control of the reconstruction filter bank is concerned, I personally can't imagine anything short of a computer being able to get close to the agility and speed required to give useful results.

OK, flame away

We still need to know the CV range for this thing!

[L-1]

We still need to know the CV range for this thing!

I did not count on CV processing, only on interpatching and made CV ins-outs before VCAs. Range 6.1mV/Db. Some explanation here between pages 3 and 4.
http://www.thatcorp.com/datashts/dn02.pdf

I think better for processing was to made CV out just after level detector.