The HP 8600A Problem has Returned

[Kevin Kuehn]

Originally Posted by Chris K

I’ll get on that and back to all of you midday tomorrow. I don’t know how much drift from the 0.56 ohm the regulator function tolerates and my comment earlier that it is “spot on” isn’t exactly accurate. My test leads have around 0.3 ohms of internal resistance and I’m getting around 0.9 ohms when I measure it so accuracy to within 0.01 ohms isn’t something my DMM can do.

Chris,

If you by chance have an accurate reading ESR meter, that would be ideal for reading such a low value resistor value. The one I have allows you to zero out the test leads resistance.

[Chris K]

I think the resistor is a good place to start. I saw one on Mouser at 5%

[Chris K]

Got a 2 watt wirewound at 1% for $6. Im not sure the resistor currently on the board is a wirewound resistor. It looks carbon comp to me

[Kevin Kuehn]

Originally Posted by Chris K

Im not sure the resistor currently on the board is a wirewound resistor. It looks carbon comp to me

I thought the same.

[Penthode]

Looking at the circuit, I would not substitute a 7805. I do not think the 723 is necessarily at fault. The circuit is very simple and I am curious if you have performed a few basic tests?

First did you confirm if it may be going into current limiting mode? What would help is when the voltage starts to droop is to measure the voltage at the regulator transistor and put the readings on the diagram.

I suspect the transistor or components around ity may be bad.

[Penthode]

Reading above it looks af if you are going in the direction of current limiting..
.

[Kevin Kuehn]

Originally Posted by Penthode

Reading above it looks af if you are going in the direction of current limiting..
.

I have a hunch, but I'm not an electrical engineer. Simple circuit, but looking at the formulas to calculate resistor values to limit current makes my head spin. Would also be good to look at the unregulated DC coming into the regulators with a scope. Check it for ripple.

[Chris K]

The power transistor is a 2N3055 and it is mounted on the chassis next to the regulator board. I have not done any testing other than looking at the voltage on the test point and observing the symptom. The construction of this instrument makes doing some of the suggested testing very difficult. I imagine if I study the schematic a bit more there would be places to measure the voltage easier than directly on the 3055 which would entail removing the bottom of the unit. If I understood what happens in the regulator when I use the freeze spray and it restores function for a while until it gets back to operating temperature I might be able to help figure this out. Why does freezing the 723 make it regulate properly?

[Chris K]

Penthode, I did think measuring unregulated voltage into the 723 vs the voltage out would be useful but with the board plugged in, it’s going to be physically difficult and I don’t want to buy an extension for the connection slot just yet. Makes me appreciate point to point wiring in tube gear!

[Penthode]

I am an electronics engineer and have designed regulated power supplies using the 723 for many decades. I wanted to see voltage readings around the 723. The 723 is an old and reliable device.

I think this will be a simple problem to resolve. Why do you not first measure the voltage directly across the 0.56 ohm resistor as this will indicate the current from the +5v supply and immediately identify if the 723 is going into current limit mode.

[Penthode]

I am an electronics engineer and have designed regulated power supplies using the 723 since the '70s. I wanted to see voltage readings around the 723. The 723 is an old and reliable device.

I think this will be a simple problem to resolve. Why do you not first measure the voltage directly across the 0.56 ohm resistor as this will indicate the current from the +5v supply and immediately identify if the 723 is going into current limit mode.

[Kevin Kuehn]

You can tack solder extension wires onto any circuit board points you want to measure with the board plugged in. Such as across the current sense resistor.

[Penthode]

The two connections across the 0.56 ohm resistor are current limit and current sense. When the voltage across the resistor reaches 0.7 volts, the current limiting begins at the output voltage will begin to decrease if the current is exceeded.

You need to accurately measure the 0.56 ohm resistor. If it has increased even slightly, you will see the problem you are exhibiting. But measure the voltage across it first for if the resistance increases, the voltage across it will increase. Right now it is designed to limit at just over 1 amp.

If the voltage across the resistor exceeds 0.6 volts, there is one of two problems: the resistor has increased in value and needs to be replaced or there is a problem with the circuitry downstream eg a leaky capacitor.

If the voltage remains below the 0.6v and the output voltage drrops, then there is a problem with the 723 circuit. If you have replaced the 723 with a known good one, I cannot see that there is a problem there.

[Kevin Kuehn]

Originally Posted by Penthode

Right now it is designed to limit at just over 1 amp.

Can you predict how much headroom they would design in, for example 500ma, or would they run much closer to the limit?

[Penthode]

Originally Posted by Kevin Kuehn

Can you predict how much headroom they would design in, for example 500ma, or would they run much closer to the limit?

There should be a knee at just over 1 amp. So the voltage should stay at 5.0 volts up to the knee. I would suggest 0.6v is a good maximum where the roll off begins. In normal operation, it should be running with no more than 0.4 v or so across the resistor