Casio DG-20, power supply

Guess what I bought last weekend.. :D

The instrument Darth Vader would use if he was into playing..

Yes, it’s the silly looking (and ridiculously overweight) guitar synth straight from the 80s, made famous by Flight of the Conchords.. the Casio DG-20! You probably now know what to expect to be reading in this blog over the coming weeks.. ;)

I got my DG-20 from Wolf Rami of Nightsatan. Apparently he had bought a few of these and used them to piece together a single working one for his own use.  So I knew exactly what I was getting into and actually bought the unit, if no good for anything else, just for the enclosure and string mech alone. Cheap “raw materials” and all that.. Should the synth indeed turn out unrepairable, alternative uses shouldn’t prove to difficult to come up with. Anyway, Wolf warned beforehand that the electronics have presumably seen a bit of overvoltage and that the battery compartment cover and one headpiece screw are missing.

Arrival check and indeed, the synth doesn’t power up. Lucky me, Dan Newcome has been kind enough to upload the service manual online. This helps a lot with analysing the fault, as there’s no need go about it the hard way and spend (a lot of) time reverse-engineering the power circuit. With the help of the manual, I was actually able to analyze possible problem areas before I even got my hands on the synth!

Schematic for the power supply circuit of DG-20

If you look at the picture above, there are two main subsections in the power supply circuit. When divided vertically in half, there’s the “switch on” section formed by transistors T1-T3 on the left, and the “output” section formed by transistors T4-T6 on the right. Quoting the service manual, the “power on” sequence goes like this:

1. When the power switch is turned on, battery voltage is provided to the base of transistor T3 causing the transistor to turn on. Current from T3 emitter charges the 10µF capacitor.
2. The electric charge from the capacitor turns T2 on compulsorily.
3. Transistor T1 turns on providing the battery voltage to voltage regulating transistors T4, T5 and T6.
4. After being charged, the capacitor does not provide voltage to T2 base however, receiving VDD (+5V), the CPU provides +5V to T2 base by signal APO.
5. If the guitar is not played for six minutes, CPU lowers signal APO causing the transistor T1, T2, T4, T5, and T6 off.

..so in other words, the circuit also has a simple “battery saving function”, controlled by the signal ‘APO’ . In case you’re wondering “APO wtf?!”, this is probably derived from something like ‘Auto Power Off’ and not eg. Asian Productivity Organization..

..

Briefly checking for internal operating voltage, the +5V bus measured clean zero. Luckily no short-circuit to GND and no input voltage on the “output” section either, so the fault was now pretty much narrowed down to “switch on”. Checking for the functionality of this subsection step by step, it immediately turned out that T3 isn’t doing much. Here, the voltage “visible” to the base of T2 measured about 0.2V. Just to verify that there isn’t a connection issue between transistors T2 and T3 (they’re on separate boards), I tried to bend T3 back slightly to get a better contact for the multimeter lead. When doing this, a piece from the T3 shell fell off! Take a wild guess which is more likely cause of problem, a damaged T3 or a connection issue between the two transistors ;)

Pretty obvious damage on transistor T3..

The service manual lists T3 part type as 2sc1310F-T or 2SC1740LNSR NPN. Both seemed a bit obscure to be sourced from anywhere (read, very little plausible search results) so I didn’t bother looking deeper into compatible replacements. 2N3904 I have a plenty of, so I figured I’ll just give that a spin. It’s not like the part is doing more than off/on switching, so anything remotely similar should work just fine. Come to think of it, you could also bypass the whole “switch on” subsection, if the APO function is not needed.. Comparing spec-wise, the 2N3904 isn’t exactly a value-for-value match with the 2Sc1730. However it does have higher maximum values on all common ratings (Vceo, Vcbo etc.).

Other than this, 2N3904 isn’t pin compatible either. The base is in the middle pin, whereas on the 2SC1730 this is on the right pin.. A bit of plier work needed, so to say :)

2N3904 viewed from both sides, the bent base terminal is more visible on the flipside.

And that’s about it, synth powers up once again \o/. Gotta admit, the internal sounds are pretty (horrible) bog-standard “Casio 80s keyboard style”, but definitely have to test running them through guitar pedals and such later on ;)

To keep this post short, I’ll save the mech stuff for later. Dan has a bit of DG-20 disassembly pics in his blog post here, in case you want to check out a bit of the mech stuff beforehand. To boot with the situation is this for the headpiece screw:

Side-by-side, a complete and a damaged headpiece screw

I’m thinking a screw and a lathed board riser should work here just fine, but let’s see what I come up with.. Luckily these parts are all metric, so I can get away with M3 hardware I have plenty of. No need for silly imperial body part dimensions ;)