Well, as the subject says, if I turn filter tracking on (aka Keyboard Control), oscillator tracking turns flat, quite a bit. Keyboard Control 2 has the biggest effect, but Keyboard Control 1 contributes too.
I’ve calibrated the PSU and the oscillators using the procedure outlined here and as long as I keep Keyboard Control 1 & 2 off, oscillator tracking is fine.
I’ve also calibrated the filter’s tracking (it was waaay off when I received the synth) and it also tracks fine, with Keyboard Control 1 & 2 on.
Did I miss some calibration procedure, or is there something wrong? Where should I look?
I must confess that I didn’t measure the PSU voltage at the oscillator board, but rather at the PSU’s heatsinks. Got that from a post by Keving Lightner though, so I figured that would be fine?
BTW, the keyboard was cleaned and had all bushings replaced buy the seller’s tech. Feels great and is really silent. But could an error when the keyboard was reassembled account for my issue?
Sounds like a bad contact on one of the board connectors. Possibly pin 4 and 5 of C02 (connector of board #2). Either that or possibly a cracked solder joint on one of the connectors wires. Maybe on pin 6A of board #1 ? Or it could simply be a bad contact in the Cinch connector for the keyboard under the mod wheels assembly …
Anyway, it sounds like the current is dropping on the pitch CV when assigning it to offset the filter cutoff (with Keyboard Control switches engaged). A sure sign of poor contact or loose connection somewhere.
Either that or get out the magnifying glass to look for cracked solder joints.
I would first try disconnecting and reconnecting the keyboard connector (Bakelite Cinch connector notorious for oxidized aluminum pins). Accessed easily by removing the mod wheel assembly (two screws on top front).
While you’re there, try to clean the Mod Wheels connector also (same type of connector).
It’s s/n 7800 so it’s not very early. It’s been “upgraded” with the “new” (third gen) oscillator board, but I’m not sure if they swapped to the newest buffer board at the same time.
But could a faulty (or mis-matched) buffer board account for the behavior I described? The oscillators track fine throughout the octaves, just as long as I don’t turn on filter tracking.
If someone has simply dropped in a newer/later oscillator board to replace the original inside your Mini, without removing/disconnecting resistor R59 on switch #3 (range of oscillator 1), it may cause some pitch control issues.
BTW, all Minis (so equipped) have the same buffer board though. And yours being in the 7800 certainly has it already installed at the factory.
OK, according to the Lords of the Mini site, a new buffer board accompanied the third gen oscillator board, at least that’s how I understand their text. But that site’s info’s been questioned before as I understand it?
I certainly have a buffer board. And R59 is removed.
I’ve cleaned the mod wheel assembly and keyboard connectors and resoldered a couple of bad joints. I’ve also cleaned the edge connectors of the oscillator board and board #2 (already cleaned other boards’ connectors) with isopropanol. Unfortunately, it didn’t make any difference (though I’m sure it will in the long run).
However, I found a small breadboard in a plastic bag connected to the front panel glide pot and to the keyboard circuit part of board #2. The board has an 8 pin DIL (an op-amp?), a small electrolyte cap and a couple of resistors on it. No trimmer.
Also, couple of wires coming from the keyboard or mod wheel assembly have been cut
And Q24 and Q14 are soldered to the back of board #2, so they’re possibly replaced with different values.
Any idea what this mod might be about?
The only odd thing I’ve noticed about the mini (except the tracking issue) is a very small modulation bleed and that the mod wheel seems to be a bit extreme - it’s really difficult to dial in subtle modulations. No idea if these things are in any way related.
(Lousy joints, dirty bars, bad connections and alcohol you really shouldn’t drink - restoring a mini is just a runaway spouse from being a country song, isn’t it… )
Transistors soldered on the traces side of a pcb, a breadboard enclosed in a plastic bag connected to the glide pot ? Never seen a mod like that. don’t know what it is. It could be a mod to get an exponential glide instead of the linear one provided originally ? I don’t know. But since this contraption is connected to the keyboard pitch circuit, it is very possible that your power drainage of the pitch CV when flipping the Keyboard Control switches in the filter section might be related to it ? Again, without knowing exactly what this mod does and exactly where it’s connected, I can’t tell for sure.
As for the mod wheel, there will usually be a very faint bleed of modulation even when the wheel is at minimum position. That’s normal for a Mini. The wheel works like a water tap, and reduces the modulation signal when closed. But there will always be a small amount of bleed through due to the way the Allen Bradley potentiometers are made. There is always a very low resistance value when they’re closed because the wiper inside is made up of a carbon pad (much like a motor brush of an electric motor). This offers decades, if not centuries, of longevity but not 0 ohm measurement when closed. That, I’m sure of.
Yeah, I remember reading about the bleed thru, now that you say it. Doesn’t really bother me.
The mystery mod does though and it sure seems like a likely cause of the tracking issue.
I’ve figured out as much as that it taps the -10v/+10v/ground pins, the pin out to oscillator and filter tracking - and to the source pin of the Q13 FET. This is the point where the signal from the glide control normally would enter.
Q24 is replaced with a 2N3904 (originally a 2N3392). Q14 is replaced with 2N3906 (originally a 2N4058).
The severed cables are the ones that used to go to the glide pot.
I’m very tempted to remove the mod, but the substituted transistors worry me. I’m an electronics n00b, so when comparing the specs with the originals’, all I can tell is they’re pretty different, but not if they’re different in any way that matters?
I decided to remove the mystery mod and see what happened. I also noticed that the keyboard pitch wire going to board #2 had been cut, then soldered together again, poorly and with no insulation, so I decided to resolder it.
I left the glide pot disconnected, just in order to avoid an extra cause of issues. Plugged the boards in, turned on the mini and now the keyboard played the same note over the entire range. Disconnected the mod wheel assembly, and now all I got was one very, very low frequency drone.
So I suspected the resoldered wire, but nope. Then I inspected the Cinch Jones-connectors and though I did notice a few modifications to the wiring of the mod wheel assembly, they surely couldn’t account for the drone. So I looked at the board again - maybe I managed to drip some solder when I desoldered the mod? Nope. But a few millimeters from where one of the mod’s wires was attached, the PCB trace for the -10 v line was cut, apparently on purpose.
So I mended the trace and voila - now the keyboard works correctly. And voila! Turning filter tracking on, no longer affects oscillator tracking!
I’m gonna have to recalibrate the oscillators, for the second time this week, but that seems like a small price to pay.
I’ve yet to reconnect the glide pot, but that’ll have to wait till tomorrow. Haven’t noticed any adverse effects from the replaced transistors, so I’ll leave them in.
Sounds like your Mini had suffered from a bad hack job by unqualified/unprofessional tinkerers !
Luckily you seem to have succeeded in getting it back in shape. Good Job !
I don’t expect any problems in reconnecting the glide pot. Unless the bad mod has modified some traces in the glide circuit too.
Hopefully not.
Yeah, it feels great to have gone through all this. I’ve recapped the PSU and rectifier boards, cleaned all edge connectors, fixed a dead headphone out, calibrated the PSU, VCO’s and filter and fixed the tracking issue. And in the process, I’ve read the service manual and Kevin Lightners collection of tech docs back and forth and I’ve read just about every post I’ve found here and on the interwebz about Model D restorations. So it’s been a very educational and gratifying experience.
Of course, your support and help have played a big part in this, so thank you sir! If you ever find yourself in Stockholm, Sweden, I’ll buy you a fine beer or two. We have some really nice micro breweries around here.
Hm, I’ve reconnected the glide pot now and the glide is acting up.
Glide down is considerably slower than glide up. Is this by design or a problem with my mini?
If i set the glide pot to > 7, it turns into a pitch wheel… The higher I turn it, they higher the pitch.
I guess the latter behavior could be explained by the replaced transistors. But if I read the schematics correctly, the glide circuit consists of Q10 and Q13 (and C6 and a couple of resistors) and they’re not replaced.
On a properly working glide you should get maximum glide times around 2.6 seconds (up from lowest note to highest note), and around 4.2 seconds (down from highest note to lowest note), timed with a stop watch, when glide knob is set to 10 (maximum). The small discrepancy between glide times is normal, it is by design. I remember that there is a modification around on the web somewhere to have them be the same though.
As for your glide issues, since there are no transistors added on the traces side on the original Mini, I would remove the ones you have. Maybe that will fix the glide issues ?
EDIT: Since these transistors seem to be connected on the same traces as transistors on the component side (identified by their pinout E B C, I would make sure that the original ones are still there and that the ones added on the traces side are not simply replacing missing original ones ?
Also check that the glide potentiometer is properly connected (wires conform to schematics) for proper operation.
According to schematics, you should have a violet wire connected to center leg, and a blue one connected to one side. They should both go to connector C02 pins 8 and 7 respectively.
The transistors on the traces side replace the original transistors, so I don’t have the originals.
Q24 is replaced with a 2N3904 (originally a 2N4058). Q14 is replaced with 2N3906 (originally a 2N3392). (BTW, I got the original’s values the other way around, earlier in the thread.)
What’s more, Q14 should be part of a matched pair, together with Q23 (which retains its original value), so that one’s really suspect.
It’s just that it Q14 and Q24 doesn’t seem to be part of the glide circuit as far as I can tell, but part of the S/H. But I’m an electronics n00b, so I’m certainly not sure! And they’re all part of the keyboard circuit of course.
The wiring conforms to the schematics (it didn’t before I removed the mod and rewired).
connected to the front panel glide pot and to the keyboard circuit part of board 
)
