Centipede question

[bit_slicer]

Very true. to test out your -5V you'd need to put it in a machine that accepts a -02, -03, or -06. The AR-II page at EF lists what games take which AR-II:

http://www.elektronforge.com/AudioRegs.htm

 

[ThurworAlpha]

"Bad" 7905s

What I was actually getting at was that the 7905 has a maximum input voltage of -25Volts and that it would have to dump roughly 18.8 Volts at whatever current(over 1 Amp you said) as heat on the heatsink.(18.8+ Watts!) The other thing is that the screw on the heatsink is plastic and 30 years old... You really need from 8 to 16 in-Lbs of torque on that screw and the device has to be perfectly flat against the heatsink, using heatsink compound. I always use the steel screws with the bushings myself. The max listed current capability is 1.5 Amps if it is properly heatsinked. (I have used them beyond this with a big heatsink)
So it is at the max input voltage and you measured it at the max rated current....Hmmmm.

Anywho... You have a couple of choices:
1) change out the chips that use the -5 Volts they are probably bad if they are drawing over an Amp. (on my centipede it looks like it only goes to one chip on the main board an LM324 shouldn't be over 100mA Max.

2) Reduce the input voltage to the 7905, which will reduce the heat dissipation, and keep from killing it.

3) Use a higher current rated regulator.

4) Make absolutely certain that it is installed correctly, use thermal compound, and make it tight against the heatsink.

I have a thought about reducing the input voltage... on my board Q11 is not used. I could put in the LM7915 and C26 and C30 and use it to feed the LM7905, thereby dropping the input voltage and reducing the 'dumped' voltage by approx 1/2. this would reduce the heat dissipated on the 7905 by a lot.

Just thought of it R31 is cement on my board and could be causing a problem due to age and moisture absorption, which might oly show if it is powered up.

What do you think?

 

[bit_slicer]

In my experience, if a manufacturer *specifically* says not to pull more than 1.5A from a voltage regulator, then you don't pull more than 1.5A from a voltage regulator, despite what you *think* might be the issue. Having said that, it is quite a voltage drop. And we are definitely running them at or above their limits. However, the newer LM7905 parts do have a thermal shutdown so at worst the voltage should fall over before they destroy themselves. The thing is, we've had folks who have removed the part, replace the heatsink compound, reassemble, did a continuity check, and they still fail.

Anywho... You have a couple of choices:
1) change out the chips that use the -5 Volts they are probably bad if they are drawing over an Amp. (on my centipede it looks like it only goes to one chip on the main board an LM324 shouldn't be over 100mA Max.

That's a good point. The LM324 should not be drawing that much current. There must be something else going on.

3) Use a higher current rated regulator.

That gets my vote.

4) Make absolutely certain that it is installed correctly, use thermal compound, and make it tight against the heatsink.

Again, we've had folks do this and they still fail.

I have a thought about reducing the input voltage... on my board Q11 is not used. I could put in the LM7915 and C26 and C30 and use it to feed the LM7905, thereby dropping the input voltage and reducing the 'dumped' voltage by approx 1/2. this would reduce the heat dissipated on the 7905 by a lot.

This is a good idea, but looking at the rev -02 board (the only one I have in front of me at the moment) there are no land pads for the -15V circuit. You'd have to drill a hole in the heatsink and deadbug the regulator to it.

Just thought of it R31 is cement on my board and could be causing a problem due to age and moisture absorption, which might oly show if it is powered up.

Check the input voltage to the 7905 after it's been powered for a while. Unfortunately I don't have the ability to do this at the moment.

 

[bit_slicer]

I figured out why I was getting such a high current reading on the -5V. The LM324 was bad, probably shorted internally. It was smokin hot. Replaced it, and the current went down to about 1 mA - right about where it should be.

So... back to square one.

 

[ThurworAlpha]

'Bad' 7905s

1 mA may not be enough to cause the 7905 to regulate. I have had some regulators that will not properly regulate until there is more than about 20mA. I checked out the data sheet and it looks like it needs at least 10mA to go into full regulation. (an estimate, I didn't do all the calculations because I am lazy!) If the problem that it is not regulating you would get a higher than normal reading.
I saw another thread where someone suggested putting a pre-load resistor across the output capacitor to make it always regulate. I like overkill so I would try a 50 - 100 Ohm 1 Watt or better resistor across C24.
Again I don't have a working one so you-all will have to give it a try for me.

cross your fingers!

 

[komodo]

This is a very informative thread. I can't contribute anything except to say keep up the updates. I can gladly donate ARII boards to anyone that needs a specific model to run tests. Their current (no pun intended) state is unknown.

 

[bit_slicer]

1 mA may not be enough to cause the 7905 to regulate. I have had some regulators that will not properly regulate until there is more than about 20mA. I checked out the data sheet and it looks like it needs at least 10mA to go into full regulation. (an estimate, I didn't do all the calculations because I am lazy!) If the problem that it is not regulating you would get a higher than normal reading.
I saw another thread where someone suggested putting a pre-load resistor across the output capacitor to make it always regulate. I like overkill so I would try a 50 - 100 Ohm 1 Watt or better resistor across C24.
Again I don't have a working one so you-all will have to give it a try for me.

cross your fingers!

Yeah we've been roundabout with this one too. Are you thinking of the same thread I am? - Newer regulators regulate fine (within 0.5V) even under no-load conditions. That guy (I forget who it was now) physically tested old-skool vs new-skool and the older parts failed to regulate while the newer ones did. Which datasheet are you looking at?

I'd be willing to give it a try nonetheless. I've been working on a stand-alone AR-II test bench but it will be a couple of weeks before it's done. I do have another KLOVer's AR-II that has trouble with the -5V so I can try that. I really don't think the low output current is the issue, but I don't have any better ideas either.