MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

Dark Star Chaos Cap Help
MUFF WIGGLER Forum Index -> Fractional Rack Modules  
Author Dark Star Chaos Cap Help
raisinbag
I'm about to build a DSC2000 rev 0 I have had kicking around (yes since 2000). I was wondering if anyone who has build (or anyone with more experience than I) would help me with some cap questions. BOM has 4 - .1uf caps, two are Ceramic, two are mylar, would it matter if I just used all Mylar? From what I understand it should be the same thing (but some people say that ceramic are noisier)

As well, if I used a polyester where it specs a polystyrene for the 10,000pf, Would it matter?


Last but not least, there is a cap speced at 2000pf, which is really difficult to find. I have 2.2nf (2200pf) caps on hand, and was wondering if anyone who has built this would think that it would make a difference.

Don't want to open a can of worms with a crazy cap argument.

Thanks for your help. have wanted to build this for quite a while!

Raisinbag
e-grad
First of all: Basically the DSC is a Noise Module. Other than for a VCO or VCF you should be able to compromise on the parts if necessary.

raisinbag wrote:

BOM has 4 - .1uf caps, two are Ceramic, two are mylar, would it matter if I just used all Mylar?

I’d just buy two more ceramics since I order parts every now and than anyway. Otherwise I’d use whatever is at hand.

raisinbag wrote:
As well, if I used a polyester where it specs a polystyrene for the 10,000pf, Would it matter?

I could not find a 10.000pF cap for my DSC and end up using a high quality polyester cap.

raisinbag wrote:
Last but not least, there is a cap speced at 2000pf, which is really difficult to find. I have 2.2nf (2200pf) caps on hand, and was wondering if anyone who has built this would think that it would make a difference.

Caps often have quite a tolerance thus I’d use a 2nF2 cap. According to the BOM this cap should be a polystyrene thus I’d actually use a polystyrene or a high quality cap.

Hope that helps!

BTW You might be interested to learn what Harry Bissel has written on caps:
http://search.retrosynth.com/synth-diy/search/lookit.cgi?-v9905.861
Tim Stinchcombe
raisinbag wrote:
BOM has 4 - .1uf caps, two are Ceramic, two are mylar, would it matter if I just used all Mylar? From what I understand it should be the same thing (but some people say that ceramic are noisier)
The two meant to be ceramic are C16 and C18 on the outputs of the two regulators - using mylar caps there will effectively just be 'over specing' them, so it will work just fine.

Quote:
there is a cap speced at 2000pf, which is really difficult to find. I have 2.2nf (2200pf) caps on hand, and was wondering if anyone who has built this would think that it would make a difference.
No, you won't notice - it is the timing cap for the 'noise clock', so it will run just a tad slower, but you won't be able to detect it! I built mine with a 2200p too.

Quote:
As well, if I used a polyester where it specs a polystyrene for the 10,000pf, Would it matter?
This is the only one I would be wary of substituting. C8 is the timing cap for main VCO in the big chip, and hence the reason for specing a 'decent' cap like polystyrene. Sure, it will work with a polyester, but may not be as stable (with temperature etc.) as using something better. I used a Wima FKP2, polypropylene for that:

http://www.wima.de/EN/fkp2.htm

but probably mostly because Rapid in the UK sold them (though Googling for that link suggests Mouser etc. stock them too).

Building a DSC from the board up taught me how hard it can be sourcing the right components, especially ones that would fit the PCB holes (that Wima one won't for example - I had to 'extend' the one leg), and also ones available in the UK, from a BoM drawn up for the US market. So I feel your pain!

Tim
raisinbag
Hi thanks so much, that really helped. I understand basic stuff, but didn't realize the difference the polystyrene would make for reliability.

Is polypropelene just as good? Those Wima caps seem to have a good reputation, but the article calls the Styrene the "holy grail" so that is hard to beat! I am in Canada, so mouser works, but I like to try to find local stuff when I can.

Maybe I will socket that cap until I find a super duper one!

Yes this has been a fun project to find parts for. Glad I bought the Big chip 12 years ago when I bought the PCB. Now I have to make the other 2 DSC's I have sitting here. Can't believe I had these sitting here for so long!

If you know of any cool mods I would love to know.
idiotunderground
raisinbag wrote:
If you know of any cool mods I would love to know.


there are mods listed for the DSC in the blacet mods sticky

https://www.muffwiggler.com/forum/viewtopic.php?t=211&postdays=0&posto rder=asc&start=0
raisinbag
right on! those are really cool mods. I'll have to make a bigger than usual panel to accommodate some extra jacks and switch.

I found some Polystyrene caps I can order near by, so that is great. Thanks for all the help on this.
raisinbag
Hey just thought of an idea, after seeing the mod for clock out as a vc and reading your explanation of the 2.2n cap running the clock slower, would it prove useful to have two caps that can be switched, so we would have a normal clock speed and a slower or faster clock speed? Which way would you go, faster or slower? If so how much? I did find a nice 2000pf cap so I will order this, just thought I could mod a bit more %)
Tim Stinchcombe
raisinbag wrote:
Hey just thought of an idea, after seeing the mod for clock out as a vc and reading your explanation of the 2.2n cap running the clock slower, would it prove useful to have two caps that can be switched, so we would have a normal clock speed and a slower or faster clock speed? Which way would you go, faster or slower? If so how much? I did find a nice 2000pf cap so I will order this, just thought I could mod a bit more %)
I just fired mine up and measured the (internal) noise clock output from the 4046: it varies from around 30Hz up to about 29kHz, with the 2n2 cap I have fitted as C4, and R26 the standard 10k. If you wanted to slow it down, or have finer control over the clock (perhaps at lower speeds), there is quite a lot of room for manoeuvre, changing either of these components, or even adding a resistor at pin 12 (though this would generally be more restrictive on the frequency). I'd suggest building it up and getting it going, then see if you feel like changing anything, particularly before you commit to a fixed front-panel configuration, so you can add more stuff if you want to!

Tim
raisinbag
Tim thanks for firing it up and taking those measurements. I hope I got you inspired to use it! I'll take your advice and just build it up and see what I want to do from there. Cool thing is I have two more to mess around with so maybe I will make each one a bit different wink Only place I can find those funny pots is at Digikey, so I may just skip and use normal pots and mount the PCB to panel with angle brackets and standoffs.

Did you build your DSC to Blacet specs with the right pots? If you did something different, i would love to see a pic. Soldering begins as soon as I don't have to work at night!
Tim Stinchcombe
raisinbag wrote:
Only place I can find those funny pots is at Digikey, so I may just skip and use normal pots and mount the PCB to panel with angle brackets and standoffs.
I quickly gave up the idea of using them too...
Quote:
Did you build your DSC to Blacet specs with the right pots? If you did something different, i would love to see a pic.
...and in the end used standard Alpha 16mm pots bodged onto a piece of stripboard with some pot-mount brackets, and then bolted the whole thing to the main PCB with some cut-down stand-offs - there are a few pics here which should give you the main idea of how I did it:

http://www.timstinchcombe.co.uk/index.php?pge=dsc

Tim
e-grad
Nice pic. BTW I used the very same substitute for C8.
raisinbag
thats funny! I remember looking at your site a long time ago and liking how you mounted it. Did you do the silk screen on the panel yourself? My wife does silk screening for her business and I am going to try it on a panel when I build this.

Well all I need now are those two styrene caps and I'm all set, thanks so much for your help. I have been wanting to dive into this for so long but kids and finances have been quite the road block.

I'll post when I'm done, cheers! "the bag"
Tim Stinchcombe
raisinbag wrote:
Did you do the silk screen on the panel yourself?
It is a fairly standard Front Panel Express/Schaeffer panel, engraved and with the yellow in-fill. Took me ages to decide on the colour combination, but was really pleased with the way it looked when I got it!

I look forward to hearing of a successful completion of yours.

Tim
raisinbag
Hey i'm looking at having cv on attack an release and following this mod. it is a little contradictory, says to find r34 and r57, then says to salder 56k resistor to r32 and r57. my gut says I am supposed to hook up to r34 and r57 but just making sure. I am actually working on it right now, yay!

http://www.hotrodmotm.com/ds_mods.htm
Tim Stinchcombe
raisinbag wrote:
Hey i'm looking at having cv on attack an release and following this mod. it is a little contradictory, says to find r34 and r57, then says to salder 56k resistor to r32 and r57. my gut says I am supposed to hook up to r34 and r57 but just making sure.
If doing that mod makes sense for your board, then it must be from the original run, and not the later rev A board, as mine has that fitted 'as standard'. It is simple enough: trace back from the attack (decay) pot wiper(s), which should go through a 56k to the op amp inverting input; tag one end of your new 56k to that op amp input end, and the other end is your CV in!

Tim
raisinbag
Cool thnks. Ya it is old one actually bought them in 1999 when I am guessing Blacet was changing to the rev A. Man this PCB is a dream to salder! I cant wait to get the classic synth Im saving up for. Thanks for all the help!
raisinbag
Wow I'm almost done and ready for test. Only missing a few caps (which I'll socket until proper ones arrive) using a styrene 1800pf inplace of 2000pf and polyester 10,000pf until styrene comes in.
I need opinion on the 7.5k and 91k resistors. Man what a PITA to find these close to home. If I make a series resistor "tent" using 6.8k and 1k do you think that would be OK for 7.5k.? I looked at schemo to see why a 91k and don't really understand what it is doing after the 56k to grd. So if I put two 180k in parallel and it measures close to 91k will this mess something up? I think it controlling feedback to the opamp? Will putting them parallel mess up wattage rating?
Once I deal with these silly resistors, ill create some CHAOs!
Thansk!
Tim Stinchcombe
raisinbag wrote:
I need opinion on the 7.5k and 91k resistors. Man what a PITA to find these close to home. If I make a series resistor "tent" using 6.8k and 1k do you think that would be OK for 7.5k.?
The 7k5 resistors are used to sink current out of various pins of the big chip - using a 6k8 directly will work, and I assume the 7k5 value was chosen to ensure a reasonable margin for not 'overstressing' the chip, but the datasheet would need consulting to be really, really sure (though it may not be given, the 7k5 may have been arrived at by experimentation); 6k8 + 1k will work fine, but if you want to be on the safe side, then using 8k2 resistors instead will do it, but you may lose a teeny bit at one end of the range of whatever is being controlled.
Quote:
I looked at schemo to see why a 91k and don't really understand what it is doing after the 56k to grd. So if I put two 180k in parallel and it measures close to 91k will this mess something up? I think it controlling feedback to the opamp? Will putting them parallel mess up wattage rating?
From memory, if you control the VCO in the 4046 with the pin 9 voltage, as done here, it only has effect from a volt or so above ground, to a volt or so below the rail, i.e. it only works from about 1 to 4V. The 91k + 56k resistor (the 56k isn't to ground BTW) is generating an offset sufficiently large (approx 1.1V = 5 x 33/(91+56)) so that varying the noise clock pot will always have an impact, no matter where it is. Decreasing the 91k will increase the offset, so you'd lose some lower frequencies; increasing it will decrease the offset, so you'll have to turn the noise pot a little off zero to get any change at the lower end. Two times 180k in parallel will work just dandy, as probably would a 100k instead of 91k. Whatever you do, assuming you are using standard 1/4 watt resistors, voltages and currents are so low that the wattage simply isn't a consideration!

Tim
raisinbag
Wow Tim you certainly know your stuff! I have 6.8 and 8.2, but will probably try to keep it close to 7.5 as possible so I don't loose any range.

Oops you are right it doesn't go to ground, I peeked at the schematic that was tucked inside my other documents at work, kinda like hiding a dirty magazine inside a comic book wink. Would u mind explaining your formula (this is the stuff I'm trying desperately to learn). So the voltage drop at pin 9 is the "rail v" times the feedback resistor over the series resistance to the rail? I want badly to understand how this works and am at a stage where I have so many questions, plain old reading isnt giving answers. I notice on the schematic that there is a note "3pl" not sure what that means. And there is another note in a box says r7 on revA is 110k so it looks like a 100k resistor may be fine after all! What did you use at r7?

Thanks Tim!
Tim Stinchcombe
raisinbag wrote:
Would u mind explaining your formula (this is the stuff I'm trying desperately to learn). So the voltage drop at pin 9 is the "rail v" times the feedback resistor over the series resistance to the rail? I want badly to understand how this works and am at a stage where I have so many questions, ...
The noise clock circuit driving the 4046 VCO is basically two cascaded inverting op amp configurations: the second one is just x -1, i.e. unity gain (and inverted), because the resistors (56k) are equal (my schemo U1 5/6/7 with R9, RN1 7-8); the first one with R7 in the feedback (33k my schemo) basically attenuates a little, being x33/56 for the voltage at the noise clock pot wiper, and the previous formula for the offset generated by the fixed voltage of the rail (the two get added - can't see 'op amp summer' near that Wikipedia article, but it'll be there somewhere..). Thus the range of voltage at the 4046 pin 9 is the 1.1V from the offset with 0V at the pot wiper, and when the pot is fully CW, at 5V, we add to this 5x33/56 = 2.9V, giving a max V of approx 4V at pin 9, which ties in with what I thought about the range of usable voltage at that pin (I'm sure I've seen this plotted somewhere, but none of the half-a-dozen 4046 datasheets I have seem to show it, so maybe it was in a book...)
Quote:
...plain old reading isnt giving answers.
Don't give up, plain old reading will very likely be very necessary to get you there!
Quote:
I notice on the schematic that there is a note "3pl" not sure what that means.
At a guess it could mean '3 places', for a repeated sub-circuit, or component value? I can't see that on the schemo I have, which is at 'Rev D' (what component is it near/regarding?)...
Quote:
And there is another note in a box says r7 on revA is 110k so it looks like a 100k resistor may be fine after all! What did you use at r7?
...so the circuitry may be a little different in this area between the two versions we each have...??!

Tim
raisinbag
that label is right beside the r7. obviously no big deal. I was just curious. Your explanation was helpful thanks, I won't give up. just gave my DSC a test drive. Didn't get too far! Was short a 79l05. I'll grab one where ever I can and try again.

6.8k + 1k gave me a 7.65K on the multi, not bad. going through my 100k ple I found a 95K Can't complain.

Cheers, and thanks for all your help. Be bugging you again soon.
MUFF WIGGLER Forum Index -> Fractional Rack Modules  
Page 1 of 1
Powered by phpBB © phpBB Group