Members clusterchord Posted March 23, 2007 Members Share Posted March 23, 2007 Wes, now that umention it, on which locations, while replacing all capacitors in an analog synth, such as this one, is it safe (and perhaps desirable) to go with higher capacitor values than original, that won't affect the sound of the machine ? in peticular case of obxa, i know we have to abide to exact values accompanying 3320s on filter circuits otherwise filter changes the sound. anywhere else? i imagine bigger caps on power supply might give as less noise, is this true ? excuse my lack of knowledge here.. only thing i did change when i was re-capping my obxa and mks80 is i went with 105C instead of 85C -rated caps. but havent strayed from the original capacitor values. but since im preparing to do this to SH-2, and soon to Jupiter 8, i'm wondering .. PS did u receive my last mail about them voiceboards n wood endcaps ? Link to comment Share on other sites More sharing options...
Members urbanscallywag Posted March 23, 2007 Members Share Posted March 23, 2007 in peticular case of obxa, i know we have to abide to exact values accompanying 3320s on filter circuits otherwise filter changes the sound. anywhere else? i imagine bigger caps on power supply might give as less noise, is this true ? excuse my lack of knowledge here.. For AC coupling you don't want to change the value. You may change the low frequency response of the signal. On schematics these will be in series with the signal. Decoupling power supplies is sort of black magic. Its something you've got to sit down with a scope and put a capacitor here or there and see if the noise on the rail goes down. Downsides to adding more capacitors [or increasing values] are possibly instability and overloading the supply at startup especially with some of those old weak power supplies. Maybe add a softstart circuit! Link to comment Share on other sites More sharing options...
Members urbanscallywag Posted March 23, 2007 Members Share Posted March 23, 2007 A bigger AC coupling capacitor will increase [or extend] low frequency response, so there goes that idea. Link to comment Share on other sites More sharing options...
Members gilwe Posted March 23, 2007 Author Members Share Posted March 23, 2007 Wes, is the problem on my OBXa familiar to you ? a reminder - non of the voice cards led light up when playing, and no sound is heard. Finding - a fried 6.8uf tantalum cap placed over two legs of the LM7915 (big) transistor at the back of the machine, and a dead battery... I would really appreciate any direction ! Link to comment Share on other sites More sharing options...
Members urbanscallywag Posted March 23, 2007 Members Share Posted March 23, 2007 That big transistor is a regulator. Do you have an oscilloscope or even multimeter? You need to check out that -15V rail. That tantalum is likely shorting the -15V supply to ground, so take it out (after verifying with multimeter for example). Link to comment Share on other sites More sharing options...
Members flattop Posted March 23, 2007 Members Share Posted March 23, 2007 The cap should be replaced certainly, but you need to find out if your -15 v PS is down. If not, then the cap has no bearing on the issue you are faced with. Again, I would do some quick tests. I would verify PS at the voiceboards, then check if you have the oscillators running and tracking the keyboard. If you don't, then you are dealing with an instrument that has the digital side running, but the analog side of things gone bad. This is usually the domain of the DAC or PS. If you have osc running and tracking, then it will be the world of the gates not functioning properly. Though rare that all the dipswitches gcould go bad together, it is possible. 25 years of not being used can build up a significant amount of carbon, pushing the contacts away from each other. Similar to what happens in a pot that is never turned. If it is not the dipswitch or PS, then you have an issue that will take relatively sofisticated equipment and skills. There is a down side in that once you are past the gating problem, you may discover other issues that are not apparent at this time. Hopefully, you'll have 7 working voices, but you won't know until you start triggering them. The battery has no real involvement here. You will want to reinitialize the memory when you get it up and running. You can do this with a tape or creating a patch from manual mode where you turn every pot CW and CCW, turn on and off every switch and then create a simple program that you know is yours. I usualy create an open filter, organ style envelopes, no mod, and simple octave oscillator sawtooth patch. I then save this to every location and then go back and verify. Out of range memory can do some strange things on an instrument. What you are doing here is sort of rebooting or reinitializing the memory. Wes TaggartAnalogicshttp://www.analogics.org/ Link to comment Share on other sites More sharing options...
Members MarkShovel2 Posted March 23, 2007 Members Share Posted March 23, 2007 I recall that one of the failure modes of a tantalum cap is that it shorts. That would explain the damage. So the bad news is it probably killed the power source in the process. Please review the battery replacement. The PCB looks discolored. The new battery need to be mechanically secured. I'll have to open my OB-Xa this weekend. I have been wanting to calibrate the voice tuning. Mine is a full 8 voice configuration. It looks really cool next to my OB-8. Mark Link to comment Share on other sites More sharing options...
Members gilwe Posted March 23, 2007 Author Members Share Posted March 23, 2007 Thanks all, will check and report... BTW, computer seems to work - i.e buttons responds (when no contact problems is present fo course...) Link to comment Share on other sites More sharing options...
Members Don Solaris Posted March 23, 2007 Members Share Posted March 23, 2007 For AC coupling you don't want to change the value. You may change the low frequency response of the signal. In general for AC coupling, capacitor values are actually not that critical (sometimes extending even a 1:10 ratio). For example in audio path with 10k impedance, a 33 uF capacitor will have a cutoff point at about 0.2 Hz. If you place a 4.7uF instead, you will move a cutoff point to about 1.6 Hz - nothing to worry about. In fact, you can even put a 1 uF capacitor, you will still be safe, cutoff point will be at 8 Hz, still one octave below 20 Hz. All cool, unless you are building an earthquake generating device. Link to comment Share on other sites More sharing options...
Members urbanscallywag Posted March 23, 2007 Members Share Posted March 23, 2007 You are only looking at -3dB frequency however. As the cutoff frequency approaches audible range non-linear phase response will creep in and smear the low frequency response. It may have no effect at all but at the same time it might take away low end punch. Link to comment Share on other sites More sharing options...
Members gilwe Posted March 25, 2007 Author Members Share Posted March 25, 2007 Found a couple of partially fried 1ohm resistors, at the Power Supply section, as well... So I guess it has all started here. Hope that no damage has been done to the rest of the circuit... Wes, do you recommend getting the 7915 and 2N3055 replaced as well ? I could not see any short on them... can you think of any reason why would the cap and resistors get fried ? could other circuits get damaged as result of this defect ? I already have a pair of 723 to fit instead the ones in the circuit. Thanks ! UPDATE: once replacing the two resistors, cap and two 723 ICs, powering up, the tantalum cap has got fried immediately !! any idea ?? Link to comment Share on other sites More sharing options...
Members flattop Posted March 25, 2007 Members Share Posted March 25, 2007 I would replace everything of any importance in the entire -15v/+15v line. It looks like you had a possible runaway and it took out the two *fuse* resistors and the cap. I would also replace the W02M rectifiers on all the power supply lines with W04M's. Now for the bad news. Why did this occur? You run the risk of destroying that part of the power supply again it you do not know what took it down in the first place. Before you go replacing the +/-15 volt stuff, you need to take a stab at what went wrong in the first place. Did the PS fail on its own? Did something downstream take it out? I would at least make sure you did not have any direct short on the +/-15v supply. Since you do not have a history with the instrument, is it possible that the instrument was in for service for this issue and that accounts for the missing voice board? I notice you are missing the hardware to secure voice board #6. Was other hardware missing? Symptoms of an instrument in for service and the service provider was in over his head and stuck it in a corner somewhere for a period of time. They then lost the mounting hardware and voice board over that period of time. After getting the power supply running, I would power it up without the +/-15 volt supply connected to the voice motherboards. If you get past this successfully, I would then add the motherboards one at a time. Once the motherboards are running sucessfully, I would test each voice board in the instrument individually to assure of their working condition and not taking out the power supply. Continue with each one until all have been in the instrument. If you get past this stage, then add each one to the instrument until the instrument has all 7 voices running. That said, finding a tech might be a safer way to go. I fear, at times, that I may be only loading a pistol. I worry that by giving advise, without ever seeing the patient, that I could be encouraging decisions that could make things worse in the end. Wes TaggartAnalogicshttp://www.analogics.org/ Link to comment Share on other sites More sharing options...
Members gilwe Posted March 26, 2007 Author Members Share Posted March 26, 2007 Hey Wes, thanks so much, this is what I thought I should do. As for the "fuse" resistors - are these regular resisitors which act like fuses in that particular circuit - or are they of different type (i.e "fuse resistors) ? what wattage should they be of ? I wonder what could make BOTH +15 and -15 rails go down at the same time... In case I disconnect all boards from the PS unit, and still got the same problem - could that be caused by a bad transformer ??... Can any of the diode bridges or regulator short down even if no short is observed when testing them out of the circuit ?... Thanks so much - again !! Link to comment Share on other sites More sharing options...
Members J3RK Posted March 26, 2007 Members Share Posted March 26, 2007 I'm sure this really doesn't apply, but I though I'd bring it up. A friend of mine bought a supposedly 5 voice Xpander. Turns out, all he had to do was going into a menu, turn the voice back on and calibrate it. (seems like maybe someone was having a bit of trouble with it, and instead of fixing it, just turned it off. On another one, he had to do some recapping to get one of the voice chips (and possibly another one, to work again.) We're dealing with slightly different tech of course, and I don't even know enough about the XA to know if it has the software interface that the later ones do, but just an idea. You may be able to get that board working without too much trouble. Edit:Now that I took a few minutes to read some of the later posts, I realize that most of my ideas will not help you. (except maybe the power supply parts) Link to comment Share on other sites More sharing options...
Members flattop Posted March 26, 2007 Members Share Posted March 26, 2007 Low value resistors are sometimes placed in stratigic position to be (hopefully) the fall guy when things go amiss. The theory to try to provide some protection for the more expensive parts downstream. They are just regular fuses usually. In this application, you do not want to use a higher wattage value than was originally spec'd for the job. You are faced with a situation where you can blow the PS again after fixing it or cause more damage downstream. That is why an incremental approach is needed here to try to determine what went wrong and where. Or you need to invest in a meter and determine if you have any direct shorts on either the + or - side. There is a possibility that the W02M failed in a way to destroy the parts on the PS since both the + and - sides have that part in common. I do not see transformers fail and take out only one part of the PS. Are you running 240v in your country? I've seen low voltage do horrible things to instruments. Is it possible that someone in the instrument's history turn the instrument on in the 120v setting? Another thing I would do is to remove every board from the instrument and make a visual inspection to make sure you do not have a screw or other part wedged underneath. I have found many crazy things inside instruments, even a $20 bill. Pens, loose change, screws, a condom (don't worry, still sealed - I think the instrument was his safe place from his wife), a dog tag, etc... Children seem to like to place change especially between the keys, sort of like a bank. It can get wedged in the most unlikely places. Wes TaggartAnalogicshttp://www.analogics.org/ Link to comment Share on other sites More sharing options...
Members gilwe Posted March 26, 2007 Author Members Share Posted March 26, 2007 Hey Wes, thanks so much. I was also thinking about the diode bride as the culpirt... will replace the parts in those -15 and +15 rails. The 110v/220v issue bothers me as well, will check today (we run on 220v here). So the resistors are regular 1ohm 0.25w resistors if I get you correctly ? Link to comment Share on other sites More sharing options...
Members flattop Posted March 26, 2007 Members Share Posted March 26, 2007 Yes, 1/4 watt resistor will do. I imagine the 3055 and 7915 will need to be replaced also. Wes TaggartAnalogicshttp://www.analogics.org/ Link to comment Share on other sites More sharing options...
Members gilwe Posted March 27, 2007 Author Members Share Posted March 27, 2007 Hey Wes, some news - When PS board is disconnected from the other boards, tantalum cap has not blown up ! I used a new 25v. Anyhow, voltages are odd - +15v OK (adjustable) -5v OK (adjustable) -15v reads -14.6 +5 reads +4.2 I have all parts needed to refurbish the PS - W04 new electrolytics caps: 4700uf 3300uf 2200uf 22uf (btw, schematic says 15uf for all 22uf caps used in that circuit, I also have 15uf in hand) 7915, 7805, 3005 regulators Can you think of any reason why would the +5 volatge be that much out of spec ? Might be a transformer problem ?... Link to comment Share on other sites More sharing options...
Members flattop Posted March 27, 2007 Members Share Posted March 27, 2007 Something is probably drawing it down. Is the 7805 running extremely hot? I just finished up a troubling P5 repair where my 5.6v supply was running at about 3.9v. I thought that it would be a bad ic or ram drawing it down. Ends up that it was the battery circuit allowing the 5.6v to bleed back into the battery. Never saw that before. Replaced the battery and the diode. Try lifting the diode to from the battery to take that out of the equation. Wes TaggartAnalogicshttp://www.analogics.org/ Link to comment Share on other sites More sharing options...
Members gilwe Posted March 27, 2007 Author Members Share Posted March 27, 2007 NEWS: connecting to a different ground point, voltages shown ok. I replaced all electrolytic caps, diode bridges and also the 7805 of the PS, just in case. Connected each of the synth boards to the PS, one at a time. It became that the lower control board makes the +15v rail go down to 0.4v while connected to the PS ! other boards didn't show any voltage problem on PS. It became that: 1. A tantalum capacitor marked "10 16v", and which is placed between the +15v and ground is shorted ! 2. Someone has broke the case of a part called AD7541 which I assume is a D/A converter (?) It is broken in such a way that the metal top of the IC is no more connected with leg number 1. There seems to be a trace going from the top metal area to the leg, and that connection has been broken. Both the cap and the IC are placed one near the other on the board, and there are signs showing that the cap has probably been replaced in the past... The cap reads "10 16v" but the schematics only says "6.8uf", so I guess the wrong type of cap has been placed instead of the original one ?... The schematics doesn't say "tantalum"... Is the AD7541 readily available in the market, as a standard part ? UPDATE: cap replaced with a 6.8uf tant. 35v - SOUND ! Only the 8th voice responds at this point. Continuing to check. I guess the AD is bad... Link to comment Share on other sites More sharing options...
Members gilwe Posted March 28, 2007 Author Members Share Posted March 28, 2007 OK.... so basically I can see that most functions work, contact problems here and there, nothing too serious. That beast sounds FAT ! Although it acts like a monosynth right now, I can still hear what they are all talking about ! But why do I get only the 8th voice getting triggers ... ? Is it the AD7541 ?... or would the 8th voice not work as well, has the AD7541 been defected ?.... The DIP is set correctly as far as I can tell... anything any of you can think of ? hmmm ... Calibration maybe ? It's hard to believe that all other 6 voices went out of specs is it ? Cheers and thanks all ! Link to comment Share on other sites More sharing options...
Members urbanscallywag Posted March 28, 2007 Members Share Posted March 28, 2007 AD7541-AKN AD7541-AJN I'm not sure which the OB-Xa is spec'ed at. but the AKN looks like a much better part. Better accuracy and lower temperature drift. Link to comment Share on other sites More sharing options...
Members urbanscallywag Posted March 28, 2007 Members Share Posted March 28, 2007 By the way, great job resurrecting the dead! Link to comment Share on other sites More sharing options...
Members gilwe Posted March 28, 2007 Author Members Share Posted March 28, 2007 Thanks, still working on it - and thank Wes ! btw, 4 voices are working right now. So I guess there's a chance the the other 3 voice cards are not calibrated to specs (values out of tolerance) ? Link to comment Share on other sites More sharing options...
Members flattop Posted March 28, 2007 Members Share Posted March 28, 2007 I would suggest taking the remaining 3 cards and place them in a working spot on the motherboard. If they continue to not work, you know it is local to the voicecard. Aside from that, it is a good idea to take the cards and remove them and reseat them on the motherboard. You don't have any idea how long they have sat not functioning. I would do the same for any important voice chip in a socket on the voicecards. Glad you are finding some success. My bill is in the mail. Wes Taggart Analogics http://www.analogics.org/ Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.