EDC CR103

[Spike]

Hi, all

I had been looking for an EDC CR103 for some time, for no real reason other than a fascination with precision low voltage/current circuits and I thought it might be nice to have a reference around for various projects and whatnot.  These don't come up very often, so after waiting a while with no luck, I jumped on an MV!06 on eBay for parts or repair a few months back for about 80 bucks.  It turned out that the repair was a simple broken solder joint in the divider switches and this thing was in spec. and has been rock solid ever since.

The other day a CR103 popped up on eBay for a similar price, so without much thought I grabbed that as well.  This one was also for parts or repair but I thought it might be a good education.  I should have mentioned that I'm a relative novice, but looking at the schematics I thought I could eventually get a grasp this and maybe repair if needed.

It turns out that the unit is basically functional.  That is, it powers up and short of the decimal point lamps being out, all switches are functional and it produces output that is roughly where it should be.  That said, this thing is nowhere near usable as the output has roughly 10mV of ripple at about 80KHz.

So heeding the advice that I see so often on these forums, I first checked the power supply for obvious issues and there are some.  I'd like some help in understanding how or why these are present

I figure that many folks on these forums are familiar with these old EDC units, but for those who are not the manual, with schematics, is here:

http://exodus.poly.edu/~kurt/manuals/manuals/Other/EDC%20CR103%20Operating%20%26%20Service.pdf

The schematic for the power supply differs from the unit that I have in that the two 7x15 regulators are replaced by a single TO-66 package, RC4194TK dual-tracking +-15V regulator, datasheet here:

http://datasheet.datasheetarchive.com/originals/distributors/Datasheets-11/DSA-209651.pdf

So, on to what I'm seeing.

The output, as mentioned above has 10mV of ripple (for lack of a better term) at 80KHz




The rectified, unreglated voltages are roughly +- 32V, with ripple on each the positive and negative of about 380mV at 120Hz.






This is regulated down to +-24.5V with a pair of 1N5359 zeners, which is the input for the RC4194 regulators.  The positive 24.5V input has about 8mV of ripple at 120Hz.

The negative 24.5V input, however is pretty horrible, with 400mV ripple or noise spikes, but the part that I'm not able to understand is that it's oscillating at about 80KHz.  I'm assuming that it's oscillation from the regulator, but I don't understand how that is induced.




This is translating to the regulated voltages, which are roughly +- 16.5V, with about 70mV of ripple on the positive rail, again at about 80KHz, and a huge 3.15V on the negative rail at a head-scratching 50KHz.




Now being a novice I'm not well versed on voltage regulator operation, and I'm completely unfamiliar with this ancient dual-tracking regulator, so reading the datasheet doesn't really help me understand where this 80KHz is coming from, nor whether it's an issue with compensation of the regulator or filter capacitance on the input, some combination of the two, or something else entirely.

I'm perfectly willing to step through this on my own, piece by piece, but I would very much appreciate it if someone could point me to a good starting point to understand what's happening here.  I'd hate to just start replacing capacitors or the regulator and fix the problem without understanding why it is happening.  In software we call that programming by coincidence and it drives me nuts.

Thanks in advance for any help, and please let me know if I can provide more information to help you help me.

[tautech]

Totally 

The schematic shows 2 x +- 15V supplies, both zener pre-regulated and +27V and +130V, both unregulated.

Is yours different?

[Spike]

Totally 

The schematic shows 2 x +- 15V supplies, both zener pre-regulated and +27V and +130V, both unregulated.

Is yours different?

I guess I should have been a little more thorough in my post.  There are two boards in the CR103, one for voltage and one for current.  The high voltage supply is for the current board, which I have not begun to look at in detail, though it seems okay at this point.  There are separate +-15V supplies for each board and everything I've posted refers to my voltage board.  Physically, the rectifier is on the bottom (current) board, with the +-27V (+-32V measured) supplied via wire leads to the top (voltage) board where it is pre-regulated and regulated.

EDIT: There are some high-res images of another unit in this thread, so probably no need for me to post my own.

[tautech]

The RC4194 datasheet lists 250mA is absolute max in the package you mention, that is far less than what's available from the LM7*** packages in your other unit. Can you measure the current load drawn?

[Spike]

The RC4194 datasheet lists 250mA is absolute max in the package you mention, that is far less than what's available from the LM7*** packages in your other unit. Can you measure the current load drawn?

To be clear, I don't have a unit with LM7x15 packages.  Those may have been fitted in earlier or later units but both of mine, as well as the unit in the other thread that I linked have the RC4194.   The max output current for the unit in voltage mode is specified as 50mA, so I would assume that the RC4194 would be sufficient for a properly functioning unit.

I didn't measure the current draw on the regulator as I think I'll need to desolder something for that, so I won't get to it right away.  I did, however, supply the 24V regulator input from a bench supply and the regulator itself appears to be drawing about 50mA.

Do you suspect that excessive current draw could account for the oscillation?  Can you point me to some reading material that discusses how this would induce the oscillation?

[tautech]

The RC4194 datasheet lists 250mA is absolute max in the package you mention, that is far less than what's available from the LM7*** packages in your other unit. Can you measure the current load drawn?

To be clear, I don't have a unit with LM7x15 packages.  Those may have been fitted in earlier or later units but both of mine, as well as the unit in the other thread that I linked have the RC4194.   The max output current for the unit in voltage mode is specified as 50mA, so I would assume that the RC4194 would be sufficient for a properly functioning unit.

I didn't measure the current draw on the regulator as I think I'll need to desolder something for that, so I won't get to it right away.  I did, however, supply the 24V regulator input from a bench supply and the regulator itself appears to be drawing about 50mA.

Gotcha, understood.

You could cut a track and insert an ammeter for piece of mind that draw from the reg is well within limits.

Do you suspect that excessive current draw could account for the oscillation?  Can you point me to some reading material that discusses how this would induce the oscillation?

Possibly, study P7 of the datasheet for Vout cap specs.....solid Tant recommended or electrolytic bypassed with ceramic.



Try adding some ceramics to the main smoothing caps or just replace all the 470uF caps.
It may be that they are old and buggered and therefore ripple on the Vin to the regs is more than they like.


Oh hang on....the rail with the 3.15V ripple @50 Hz, shouldn't that be at 120 Hz......faulty bridge?

[Spike]

Possibly, study P7 of the datasheet for Vout cap specs.....solid Tant recommended or electrolytic bypassed with ceramic.



Try adding some ceramics to the main smoothing caps or just replace all the 470uF caps.
It may be that they are old and buggered and therefore ripple on the Vin to the regs is more than they like.


Oh hang on....the rail with the 3.15V ripple @50 Hz, shouldn't that be at 120 Hz......faulty bridge?

That's 50 KHz

When I supplied the regulator from my bench supply, I was still seeing the 10mV@80KHz ripple on the output.  The ripple on the regulated negative rail was significantly less, on the order of 20mV@75KHz, but this still seems to translate to the output.  The positive rail seemed largely the same as before.

I'll read through the datasheet for the RC4194 and have a look at the output capacitance.

Thanks for your help so far.

[Spike]

Possibly, study P7 of the datasheet for Vout cap specs.....solid Tant recommended or electrolytic bypassed with ceramic.



Try adding some ceramics to the main smoothing caps or just replace all the 470uF caps.
It may be that they are old and buggered and therefore ripple on the Vin to the regs is more than they like.

I finally got a chance to get back to this over the weekend.  I looked at the short compensation section of the datasheet and it calls for a 1nF ceramic cap from the comp+ pin to ground on the positive side, and one from the comp- pin to the negative supply on the negative side.  It additionally calls for a 4.7uF solid tantalum from -Vout to ground, or alternatively an aluminum with a 30nF ceramic in parallel as you stated.

In looking at the board, I find that it has a 10nF polycarbonate cap on each comp pin, and only a 10uF electrolytic on +Vout and -Vout.  Placing an additional 10nF ceramic in parallel with comp- immediately made the oscillation disappear, but when I checked these caps both capacitance and ESR seemed fine. 

Moving on to the ouput caps was a different story, though.  Each of the 10uF output caps measured in the 550pF range, with an ESR in the KOhms range.  Replacing these eliminated the oscillation in the power supply and on the device output and now everything appears much closer to where it should be. Compared to my other unit the output does seem to be a tiny bit noisier, but it's really below my ability to measure it so there's not much I can do.

When it comes to passive components, capacitors are the one that can make my head hurt when I try to understand all the parameters that can affect performance in various applications.  I'm not sure why one would choose one type over another in a particular application.  Is/was it common to use film (polycarbonate in this case) caps in place of ceramics in a loop compensation application such as this one?  Perhaps they just had these in the BOM already?

Clearly the larger electrolytic works here without the ceramic high-frequency bypass but I'm not sure why they wouldn't just go with the datasheet recommendation.

Anyway, thanks, tautech, for pointing me down the right path.  Now on to the current board.