Boss PSA -120T adapter

[Bob M.]

I've got a Boss PSA-120T 9.6 Vdc regulated adapter. I've got the newer type that you can open up the case and have access to the innards via two hex screws - a great improvement on the part of Roland. I was really getting sick of drilling out those broken wall warts. Inside there's, of course, a transformer, 4x 1N4001 diodes in a bridge configuration, a couple of filtering caps, a 7809 9 volt regulator and a small (signal) diode across two of the lugs of the regulator. Does anyone know this small diode's number? What's a reasonable guess of what diode I could use in this position? Does anyone know it's orientation? I asked Boss these questions but they advised me to buy (yet another) adapter. Their service has always been the worst. It's a bit of a pet-peeve of mine, buying adapters. I have a massive box filled with broken ones (don't we all?).

Thanks,

Bob M.

[52 Bill]

What's wrong with the diode in question? When I see a diode used in conjunction with a voltage regulator chip, I assume that it is either wired across the input to output to keep the regulator from latching up or it is being used in series with the ground leg to raise the output voltage.

How is it wired in the circuit?

[Jazz P Bass]

National semi recommends a 1N4002.
Here is the datasheet.

[Bob M.]

From what I can tell, the diode in question is wired across the ground and output lugs of the 7809 regulator but I'm not sure of the orientation/polarity of the diode. It could've been installed wrong way around. I got this thing used and previously repaired and not working so I suspect that there's a chance it was wired incorrectly. I guess I could trial and error it but there are so many knowledgeable people here that I rather ask first.

Bob M.

[Enzo]

If the rectifiers are 1N400x, and this is a smaller glass one, it is most likely a 1N4148 or something equivalent.

[Jazz P Bass]

Here is a take on the 'diode across the output to ground.'
Right from the datasheet.

[Bob M.]

Yes, there are 4x IN4001 diodes in a bridge configuration. The smaller glass signal diode isn't legible (but tests good) and certainly looks like a 1N4148. The datasheet Jazz P Bass sent speaks to the polarity orientation on point. I wonder if I should use a 1N4148 or a 1N4001 (or 1N4002) and if there would be any performance or electronic difference?

Thanks for the input,

Bob M.

[Jazz P Bass]

What are we trying to 'fix'?

[Bob M.]

I'm just trying to get this PSA-120T to put out 9.xx Vdc as opposed to the 22 Vdc when I started.

[Enzo]

Your little diode is just a diode, if it checks good, it very likely IS good. Changing it to another type will have no effect on your problem.

You have a 7809 9-volt regulator IC. It works or does not. The three legs left to right are input, ground, and output. Your bridge positive corner goes to the left pin, the center pin is grounded (connected to common, assuming the adaptor lacks an earth ground) and the right pin feeds the output.

You mention the mystery diode is between two legs. WHICH TWO? If it is between the first and third, it is the reverse protection as in figure 7 on the LM7805 data sheet Jazz linked to. Unless the diode is shorted or wired backwards, that won;t cause 20v on the output.


One other possibility, though I tend to doubt this will be the case in your unit, is the diodes is wired from the center leg to ground/common. In that case the center leg of the IC is NOT ground, it is "raised" one diode drop above ground. They might do this so the 9v regulator puts out about 9.5v, to compensate for losses in the power cord and connections. Unlikely, but I mention it anyway. If that diode is open or not wired, then the whole regulator will "float" up to the supply voltage, presumably 20v.

But if you do not have this raise-diode arrangement, then you have a simple supply, and the most likely problem is a bad 7809.

[Bob M.]

I'm on my 2nd 7809 regulator. I removed all parts from the circuitboard, meticulously cleaned the pads. With all parts removed, the transformer yielded 15.4 VAC at the secondary. I replaced all caps, replaced the regulator. I tested all diodes, all tested good. The glass signal diode did turn out to be a 1N4148 (It was very difficult to read and because of this I thought it 'may' be bad). I replaced it anyway.

I was confused by the orientation/polarity of the 1N4148 diode. It is wired across the ground pin of the regulator (the middle pin) to the output pin of the regulator (the right lug). On the circuitboard, there is a (silkscreen) drawing of each diode and it's orientation. How the 1N4148 was wired when I received this adapter and how the drawing was on the circuiboard were direct opposites. The ground pin of the regulator doesn't connect to anything except one side of the 1N4148 diode and the other side of this diode connects to the negative output. I still can't get it to work. I used a low powered soldering iron and used minimal heat and solder. I have triple-tested every connection. I'm stumped! I was assuming I had a bad or lifted pad from the circuitboard trace and made sure that was not the case now.

Bob M.

[52 Bill]

The ground pin of the regulator doesn't connect to anything except one side of the 1N4148 diode and the other side of this diode connects to the negative output.

The center "ground" pin must be connected to circuit ground either directly or through a diode. If it isn't connected the regulator can't reduce the output voltage to the rated 9 volts.

I don't know if the 1N4148 diode is original to the unit or not, but if it was wired in backwards, I'd have to assume that it was installed by someone before you. It may also be that the earlier repairer also removed the ground reference to the chip. Is there any sign of a circuit board mod like a cut trace, etc.?

I would suggest that you refer to a basic schematic and add the ground reference either as a direct connection or with a diode to raise the voltage by about .6 volts dc.

[g1]

One other possibility, though I tend to doubt this will be the case in your unit, is the diodes is wired from the center leg to ground/common. In that case the center leg of the IC is NOT ground, it is "raised" one diode drop above ground. They might do this so the 9v regulator puts out about 9.5v, to compensate for losses in the power cord and connections. Unlikely, but I mention it anyway. If that diode is open or not wired, then the whole regulator will "float" up to the supply voltage, presumably 20v.

I'm less doubtful about this, as Boss makes a point of labelling the adapter specifically at 9.6V

[Enzo]

Guess I shouldn't doubt myself.