Author |
Message |
janvanvolt
Joined: Nov 24, 2005 Posts: 285 Location: Mainz, Germany
|
Posted: Sun Jul 03, 2011 5:54 pm Post subject:
Re: Ghielmetti matrices+pins on French eBay |
|
|
/mr wrote: | Perhaps some of you vocoder builders should buy these three 20x20 Ghielmetti pin matrices: |
Already got mine a while ago,but thanks for pointing out _________________ Homepage - http://www.czmok.de
My dIY - http://diy.czmok.de
Film/Music - http://gfm.me |
|
Back to top
|
|
|
/mr
Joined: Aug 05, 2007 Posts: 223 Location: Elektron City, Sweden
Audio files: 1
|
Posted: Fri Jul 15, 2011 9:48 am Post subject:
|
|
|
tomcat wrote: | /mr wrote: | Is anybody interested in getting the 1% filter caps for this project? (144+144 pieces)
I'll be ordering caps for 6 persons, likely from RS |
Possibly me. Maybe you can post the rs-numbers? |
Sorry - I'll have to take that offer back.
RS seemed short of 1% caps, so I ended up measuring thousands of cheap 5% caps instead.
|
|
Back to top
|
|
|
tomcat
Joined: Oct 14, 2005 Posts: 141 Location: earth
|
Posted: Sat Jul 16, 2011 6:12 am Post subject:
|
|
|
Have fun
I ordered the 1% caps from RS right now. They are not short of the 1% types, just short of one RS Part-No. This is the same manufacturer number as the other one, so its the same cap. Just different pricing. I assume the number posted by JH was an older stock (250+ price about half of the newer price)
0.01 166-6421
0.047 166-6465
Not cheap, 200 euros for the 1% caps .... but cheaper then farnell & co |
|
Back to top
|
|
|
/mr
Joined: Aug 05, 2007 Posts: 223 Location: Elektron City, Sweden
Audio files: 1
|
Posted: Sat Jul 16, 2011 10:02 am Post subject:
|
|
|
tomcat wrote: | Have fun |
You bet.
tomcat wrote: | Not cheap, 200 euros for the 1% caps .... but cheaper then farnell & co |
As a comparison, measuring 300 caps takes roughly an hour if you need 10 seconds for each, and some more time around it while inventing a good test routine and setting it all up - and the price for 5% caps is around 60 €. (The 5% caps seem to be made in the same process as the 1% ones, so they have the same polypropylene long-term stability - they are just spread more around the marked value.)
So for the ones of you that can't find a good source of 1% caps with a decent price and the right size (quite hard actually, most caps are too big!) it could be worthwhile to do some manual labour. It's kind of meditative actually. |
|
Back to top
|
|
|
tomcat
Joined: Oct 14, 2005 Posts: 141 Location: earth
|
Posted: Sat Jul 16, 2011 2:11 pm Post subject:
|
|
|
But you need to measure more then 300 caps
How many did you need to get 150 1% caps? |
|
Back to top
|
|
|
/mr
Joined: Aug 05, 2007 Posts: 223 Location: Elektron City, Sweden
Audio files: 1
|
Posted: Sat Jul 16, 2011 3:24 pm Post subject:
|
|
|
tomcat wrote: | But you need to measure more then 300 caps
How many did you need to get 150 1% caps? |
You don't need any caps within the 1% limits, and you don't need 150 caps with values within 1% from each other. You need 8 well-matched caps per filter, that's all. Different filters can have different values, as long as adjacent filters don't differ too much.
We've been discussing the tactics around this in the original Haible Vokoder thread. |
|
Back to top
|
|
|
tomcat
Joined: Oct 14, 2005 Posts: 141 Location: earth
|
Posted: Sun Jul 17, 2011 1:25 pm Post subject:
|
|
|
Maybe i should reread the vocoder thread |
|
Back to top
|
|
|
Cat-A-Tonic
Joined: Mar 24, 2008 Posts: 42 Location: Yokohama, Japan
|
Posted: Tue Jul 19, 2011 6:35 pm Post subject:
|
|
|
/mr wrote: |
RS seemed short of 1% caps, so I ended up measuring thousands of cheap 5% caps instead.
|
Wearing a fez while measuring so many caps must make the task much easier. |
|
Back to top
|
|
|
StephenGiles
Joined: Apr 17, 2006 Posts: 507 Location: England
|
Posted: Wed Jul 20, 2011 11:08 am Post subject:
|
|
|
I'd like a size 6 anorak please |
|
Back to top
|
|
|
zthee
Joined: Feb 20, 2008 Posts: 414 Location: Stockholm
|
Posted: Mon Jul 25, 2011 1:36 am Post subject:
|
|
|
Looking at the points that are mention in the original vocoder thread and trying to figure out what could be done.
Quote: | (1) The different replacement signals need to be balanced in level (the noise / unvoice path is way too loud, currently). It's basically a change of some resistor values, but there are differend locations where this can be done, and the best option must be found. |
So, the first things that comes to mind is reducing the value of R2645? A potential problem than might be distortion? So if it is, changing R2654 and R2649 (And perhaps R2657?)?
Or any other solution?
Quote: | (2) The noise generator is oscillating. Shouldn't be a big deal, needs to change compensation, changing one or several capacitors. |
Should we start with increasing C2630?
Quote: | (3) In my first PCB layout, one small copper trace was missing. I have already added the trace in the new layout. Normally (on other projects), I have built the whole thing again, with the new layout, before I dared selling any PCB. I don't want to do this with the vocoder. |
-
Quote: | (4) Some thresholds (like for voiced/unvoiced detection) are still untested and need fine tuning. It generally works, but a resistor value or two must be changed. |
I think I need to hear this before I actually know what the problem is
Quote: | (5) The silence bridging / filterbank part is untested. I have tested that a CV will bring up a channel volume, of course, but I never wired a bank of 21 potentiometers for this. |
Connect 20 pots - This I can do, no problem [/quote] _________________ http://www.thehumancomparator.net/ |
|
Back to top
|
|
|
tomcat
Joined: Oct 14, 2005 Posts: 141 Location: earth
|
Posted: Fri Jul 29, 2011 5:37 am Post subject:
|
|
|
Is JH on holidays or why its so silent here? |
|
Back to top
|
|
|
zthee
Joined: Feb 20, 2008 Posts: 414 Location: Stockholm
|
Posted: Thu Aug 04, 2011 6:49 am Post subject:
|
|
|
I think this build is the DIY equivalent to running Iron Man.
I'm gonna test the PSU before putting the chips in.
Here are some guesses (Based on the schematics of the FS-1A and some inspiration from other parts of the circuit. It could be a starting point?) on the resistors with no values:
RLIM = 10K?
RG0 = Jumper?
RG1 = 10R?
RG2 = Jumper?
RG3 = ?
RSUM = 5K6? _________________ http://www.thehumancomparator.net/ |
|
Back to top
|
|
|
ach_gott
Joined: Sep 09, 2008 Posts: 79 Location: Chicago
|
Posted: Fri Aug 05, 2011 8:01 am Post subject:
|
|
|
It's so beautiful.... |
|
Back to top
|
|
|
SolderSmoke
Joined: May 21, 2007 Posts: 15 Location: Sweden
|
Posted: Fri Aug 05, 2011 8:59 am Post subject:
|
|
|
Still waiting for my 1% caps. |
|
Back to top
|
|
|
zthee
Joined: Feb 20, 2008 Posts: 414 Location: Stockholm
|
Posted: Sat Aug 06, 2011 11:02 am Post subject:
|
|
|
I tested the noise section, and it works as it should. It didn't oscillate!
Luck? Probably.
This is was not powered by the on board PSU though. I'll se if it changes once I get my transformer.
While at it I tested the mic input section as well. And it works as it should. I also got a reply from JH about the resistor values - it's basically pick and choose to suit your needs.
I think I'm gonna go with the same as the FS-1A which means:
RG0 = Open (Nothing in it)
RG1 = Jumper
RG2 = 10R
And use a 5K rev. log pot as gain control.
Then we have the limiter section. I haven't quite figured it out - but I figured that putting a 22K at RG3 should make sense? That would mean unity gain if the vactrol is not active? And I guess that RLIM needs to be adjusted according to what signal levels you choose to operate within? Which means experimenting with different values...
Anyway.
I'll report back once I get more stuff going. Hopefully I'll be able to post some kind of audio next week. _________________ http://www.thehumancomparator.net/ |
|
Back to top
|
|
|
zthee
Joined: Feb 20, 2008 Posts: 414 Location: Stockholm
|
Posted: Tue Aug 16, 2011 1:17 pm Post subject:
|
|
|
Living Vocoder.mp3
Warning - There is some serious "TSCHK" and "BRRRM" in that mp3, so lower your volume!
It's still a long way to go. But it feels like a victory - it sounds like a vocoder! _________________ http://www.thehumancomparator.net/ |
|
Back to top
|
|
|
zthee
Joined: Feb 20, 2008 Posts: 414 Location: Stockholm
|
Posted: Sat Aug 20, 2011 7:38 am Post subject:
|
|
|
I've changed some resistor values and so on -
Living Vocoder 2.mp3
It sounds slightly better, yes?
Still to fix would be the voiced/unvoiced and also locating the best place to deal with the replacement signal levels. _________________ http://www.thehumancomparator.net/ |
|
Back to top
|
|
|
decaying.sine
Joined: Aug 31, 2009 Posts: 92 Location: New Haven, CT, USA
|
Posted: Sat Aug 20, 2011 6:56 pm Post subject:
|
|
|
zthee wrote: | I've changed some resistor values and so on -
Living Vocoder 2.mp3
It sounds slightly better, yes?
Still to fix would be the voiced/unvoiced and also locating the best place to deal with the replacement signal levels. |
Yes. Definitely better. |
|
Back to top
|
|
|
zthee
Joined: Feb 20, 2008 Posts: 414 Location: Stockholm
|
Posted: Tue Aug 23, 2011 2:04 am Post subject:
|
|
|
I think so too!
It's so quiet in this thread.
But I'll keep posting my findings! Hope you don't mind.
I've recorded another one, Living Vocoder Function Test.mp3 - First testing the "Tone Control" then the "Slew" and "Assymetrical Slew" functions. After that I remove the jumpers on every second band, reducing it to a 10 band vocoder, then putting them back in again.
There is some hum (and distortion from the computer). The hum is probably due to bad grounding between computer and board. It's all very sloppy connected.
The current problem I'm working on is the voiced/unvoiced section - I find it trigger to easy on transients (I.e. there's a short click in the beginning of a word, even if the beginning is a voiced). The current approach I'm trying to adjust the LPF after the 2 precision rectifiers to filter out the shorts clicks. But so far I've just manage to make it a bit too unsensitive, so it breaks up the unvoiced parts into "stuttering sections". Though I have to say that I've never used a vocoder with a voiced/unvoiced section, so I'm kinda unsure of what to expect of it! Though in my imagination if I say "T" the replacement signal should be noice, and if I say "B" it should be the excitation signal? _________________ http://www.thehumancomparator.net/ Last edited by zthee on Tue Aug 23, 2011 9:06 am; edited 1 time in total |
|
Back to top
|
|
|
tomcat
Joined: Oct 14, 2005 Posts: 141 Location: earth
|
Posted: Tue Aug 23, 2011 9:00 am Post subject:
|
|
|
Yes ist quite. Questions were asked, no one answered. Especially not the father of this project. Instead you can read posts that he answered something by email. Whatever, time will tell....
P.S. nice you came that far already zthee |
|
Back to top
|
|
|
JoeMorris
Joined: Apr 26, 2009 Posts: 161 Location: Brighton
|
Posted: Thu Sep 15, 2011 10:18 am Post subject:
|
|
|
Anyone managed to get anywhere further with these?
Desperate to hear one. |
|
Back to top
|
|
|
tomcat
Joined: Oct 14, 2005 Posts: 141 Location: earth
|
Posted: Thu Sep 15, 2011 3:09 pm Post subject:
|
|
|
Not me. Im very disappointed how this project is undersupported by Jürgen... |
|
Back to top
|
|
|
JoeMorris
Joined: Apr 26, 2009 Posts: 161 Location: Brighton
|
Posted: Thu Sep 15, 2011 3:44 pm Post subject:
|
|
|
To be fair he did kinda warn that he wouldnt be having much to do with it. He seems to have disappeared completely though which is a shame. |
|
Back to top
|
|
|
jhaible
Joined: May 25, 2007 Posts: 2014 Location: Germany
Audio files: 24
|
Posted: Fri Sep 16, 2011 1:20 pm Post subject:
|
|
|
tomcat wrote: | Not me. Im very disappointed how this project is undersupported by Jürgen... |
Well, I'm a bit surprised to read this, and quite disappointed too, by such a reaction.
I thought I had been very clear about this from the very start.
JH., 01.05.2011 wrote: |
Serenadi wrote: "I suppose you finish your own vocoder."
No. It's very, very unlikely.
I have sold my complete studio, and I don't think I have any need for any vocoder anymore, personally.
[...]
I need about 10 people who are willing to buy a PCB for 300 Euros (+shipping), in the current state. Then you can do the fine tuning described above, and you don't have to wait for me to pick up the project again, who knows when. And you could publish the corrected / optimised component values. I don't expect any change in the PCB layout to be required afterwards, just maybe 10 or 15 component values to be altered. So probably there is no need of a new silkscreen, just a table of corrections somewhere on the web. - Then, after those 10 courageous builders have (hopefully) built a working vocoder, I will make more PCBs and sell them for the same price.
What do you think? Of course not everybody of the 10 pioneers need to do the fine tuning, but there should at least be 2 or 3 of them who really will.
All of this should be administered without me (put up a web page to publish the altered component values instead of playing the back to me), but I would try to answer questions and give hints as to what could be changed, of course. (But I would like to avoid soldering and cross-checking on my prototype.)
If this sounds stupid, don't let me know.
I know it's stupid from a business point of view. But I think it might make sense from a DIY community point of view, and help this thing finally going where it belongs: to you.
|
JH. _________________ "I tell you the truth, if anyone says to this mountain, 'Go, throw yourself into the sea,' and does not doubt in his heart but believes that what he says will happen, it will be done for him. Therefore I tell you, whatever you ask for in prayer, believe that you have received it, and it will be yours." (Mk 11,23f) |
|
Back to top
|
|
|
jaidee
Joined: Oct 28, 2007 Posts: 44 Location: Newcastle upon Tyne, UK
|
Posted: Sat Sep 17, 2011 4:39 am Post subject:
|
|
|
I don't think Juergen could have made his position any clearer.
He made plain the enormity of the project ("it's a project at the border of insanity"), the fact that at the moment he has other priorties in his life, and that the ten brave souls who bought the PCBs would have to be self-supporting.
Rhetorical questions about whether Juergen is on holiday must be galling for him when he's posted a parallel thread titled "Haible Musikelektronik closed for holidays".
I'm grateful that Juergen has shared so many great projects, and frankly I've been amazed at the high level of support he's provided for all of us who've pestered him with hundreds if not thousands of questions.
Juergen has not closed the door to the possibility of some more projects in the future.
I hope that the ten pioneers can work together to bring their vocoders to fruition; but please don't give Juergen a hard time if you are individually or collectively struggling. |
|
Back to top
|
|
|
|