Topics

TM50X DC ripple?


Michael W. Lynch
 

I have TM506 that I have had for some time.

I have a PG506 that works but it shows terrible distortion on the top and bottom of the square wave when installed in any compartment the TM506. Just as a test, I inserted the PG506 into both the TM504 and TM501 and it works perfectly, the distortion is gone.

I am attributing this to excess ripple in the DC Supplies of the since the 506 shows a 500-900mV @ 120Hz (unloaded). The distortion on the PG506 square wave looks like a miniature version of the ripple waveform that you read in the TM506. The TM504 and 501 both check with much less ripple that the 506.

Am I diagnosing this correctly or is my observation and reasoning faulty?

The service manuals do not show a specification for ripple that I can find. Have I just missed it or is the spec just not there? What would be an acceptable level of ripple?

Thanks!
--
Michael Lynch
Dardanelle, AR


robeughaas@...
 

In going through the vintageTEK Museum's TM500 power modules, the most common problem is filter capacitors, which produces excess ripple. The supplies should have less than a volt of ripple with a 1 amp load. Which supply shows the ripple you quote unloaded? Does the PG506 behave differently in the rightmost (high-power) slot than the others?

--
Bob Haas


Harvey White
 

You may have missed the ripple specs, but for any power supply, especially when we're looking at your results, suggests that the ripple on the TM506 is excessive.  A bit of logic suggests that the ripple on the TM504 *is* correct, or acceptable.  Have you compared them?

Your diagnosis seems reasonable.

Harvey

On 1/29/2021 9:04 PM, Michael W. Lynch via groups.io wrote:
I have TM506 that I have had for some time.

I have a PG506 that works but it shows terrible distortion on the top and bottom of the square wave when installed in any compartment the TM506. Just as a test, I inserted the PG506 into both the TM504 and TM501 and it works perfectly, the distortion is gone.

I am attributing this to excess ripple in the DC Supplies of the since the 506 shows a 500-900mV @ 120Hz (unloaded). The distortion on the PG506 square wave looks like a miniature version of the ripple waveform that you read in the TM506. The TM504 and 501 both check with much less ripple that the 506.

Am I diagnosing this correctly or is my observation and reasoning faulty?

The service manuals do not show a specification for ripple that I can find. Have I just missed it or is the spec just not there? What would be an acceptable level of ripple?

Thanks!


Michael W. Lynch
 

Bob Haas wrote:
Which supply shows the ripple you quote unloaded?
Bob,

I did some more tests:

+33.5 and - 33.5 are shared across all 6 compartments. The +33.5 shows 400mV and the -33.5 shows 680mV (unloaded). Since the PG506 does not utilize these supplies, this ripple may be unacceptable, but is not causing the problem.

OTOH

The +11.5 supply is split with the low power compartments showing 680mV (unloaded) ripple across all 5 compartments while the High Power +11.5 shows about 62mV (unloaded). From this I can say that the ripple on 11.5V of the first 5 compartments is probably NOT good. Since PG506 derives its 5V power from the +11.5V supply of the TM50X module, this appears to be the source of the malfunction when it is installed in those compartments. The low power compartments are where the PG506 exhibited the malfunction.

Does the PG506 behave differently in the rightmost (high-power) slot than the others?
I went back and installed the PG506 in the HP compartment, it worked properly there as well as all positions on the TM504.

For reference my TM 504 tested as follows:
+33.5V @ 62mV ripple / -33.5V @ 45mV ripple / +11.5V LP @ 38mV ripple / +11.5V HP @ 30mV ripple. (all supplies unloaded). These are values that are more inline with what I would have expected from a decent linear power supply.

Harvey White wrote:
You may have missed the ripple specs,
Harvey,

I went back and did a search in the PDF for the word "ripple" in the TM506 and TM504 Manuals. While I could find many other words using search, the word "ripple" was not one of them. I also (for the 3rd time) looked very closely at these manuals, not saying I still could not have missed the specs, but If they are in there, they are well hidden. Perhaps other versions of these manuals make reference to this specification? It would not be the first time that TEK manuals omitted important information or contained an error.

In Conclusion:

I would say that, compared to my TM504, the TM 506 has issues with excess ripple in almost all the DC supplies. I am going to replace the caps in this unit, this should help the unit perform better. Until better specs are available, I will use the TM504 measurements as my "benchmark".

Thanks for all the feedback.
--
Michael Lynch
Dardanelle, AR


Michael W. Lynch
 

Followup after bulk filter cap replacement.

I replace all 7 Bulk filter caps in the TM506 with modern caps and the little PC board adapters. I used 105C 3000 hour rated NICHICON LGU and LGY caps 4 - 5600uF 63V and 3 - 22000 uF 25V.

Testing the ripple after repairs I get:

+33.5V @ 23 mV ripple / -33.5V @ 33mV ripple / +11.5V LP @ 18 mV ripple / +11.5V HP @ 20mV ripple.

This seems to indicate that even my TM504 caps are not as "good" as the ones that I installed here. Perhaps I will need to consider replacing those caps at some point?

Thanks to all who commented and offered advice.
--
Michael Lynch
Dardanelle, AR


Eric
 

I recently had to do a 5006 for the same issue. In my cause it was causing
frequency bounce on a sg503. It was really hard to keep the 503 stable.

Eric

On Thu, Feb 11, 2021, 10:56 AM Michael W. Lynch via groups.io <mlynch003=
yahoo.com@groups.io> wrote:

Followup after bulk filter cap replacement.

I replace all 7 Bulk filter caps in the TM506 with modern caps and the
little PC board adapters. I used 105C 3000 hour rated NICHICON LGU and LGY
caps 4 - 5600uF 63V and 3 - 22000 uF 25V.

Testing the ripple after repairs I get:

+33.5V @ 23 mV ripple / -33.5V @ 33mV ripple / +11.5V LP @ 18 mV ripple /
+11.5V HP @ 20mV ripple.

This seems to indicate that even my TM504 caps are not as "good" as the
ones that I installed here. Perhaps I will need to consider replacing
those caps at some point?

Thanks to all who commented and offered advice.
--
Michael Lynch
Dardanelle, AR






Roy Thistle
 

On Sat, Jan 30, 2021 at 09:01 AM, Michael W. Lynch wrote:


I went back and did a search in the PDF for the word "ripple" in the TM506
In the,
Tektronix(1974)TM 506/RTM 506 Power Module Instruction Manual(070-1786-02), Table 1-1 gives the PARD as <= 2.5 Vp-p.
Probably, you know,
But supposedly, the term PARD (Periodic and Random Deviation) replaces the older terms ripple and noise.
PARD is the Periodic and Random Deviation of the DC output voltage (constant voltage supply) from its average value, over the specified bandwidth (all other parameters maintained constant.)
It's supposed to take account of relatively high noise spikes in "ripple."
Since Tek doesn't give the specification of the "ripple"... perhaps, take the ripple as a 1/10 of the given PARD as the best case... and 1/5 of the PARD as the worst?


Jared Cabot
 

Allowable ripple is specified in the Tek TM500 mainframe tester manual, shown in the tables.

https://w140.com/tekwiki/wiki/067-1201-99


Roy Thistle
 

On Thu, Feb 11, 2021 at 11:19 AM, Jared Cabot wrote:


Allowable ripple is specified in the Tek TM500 mainframe tester manual, shown
in the tables.
Yes... but the following warning appears whenever the give them: "These voltages are quite dependent on line voltage and are sensitive to line voltage waveform distortion."
So are they really "specified?"
Probably why the specs use PARD in the actual TM506 instruction manual?


Michael W. Lynch
 

On Thu, Feb 11, 2021 at 01:19 PM, Jared Cabot wrote:


Allowable ripple is specified in the Tek TM500 mainframe tester manual, shown
in the tables.
Jared,

Thanks for pointing that out. It would have been nice if they had put this information in the Service manual for the TM506, as well as in that tester manual.. In any case, my issue is resolved for the moment. I will check this again when I get the TM50X tester plug in built. I still have a TM504 and 503 that need to be evaluated.
--
Michael Lynch
Dardanelle, AR


Ed Breya
 

The no-load, or light-load mV-ish ripple measurements on the raw DC supplies don't mean much unless you consider the possible conditions and worst-case scenario, although they may be an acceptable indicator of the caps being OK or not. If you're designing something to work reliably over the full range of possible conditions, it's good to use the TM5000 specs. The worst-case bottom of ripple is at highest ambient temperature, lowest line voltage, lowest line frequency, and maximum load. The maximum load is difficult to define. We used 8 V and 26 V to be slightly above worst-case TM500 BOR, as the minimum DC outputs in the TM5000 mainframes. That represents the minimum overhead you'll have for linear regulators. In reality, you likely won't reach worst-case, but it's good to be aware of it.

So, don't bother with no-load ripple - put a bunch of plug-ins in there to get a better idea of what to expect.

Ed


Ed Breya
 

Oops - forgot a sentence at the end: When it's sufficiently loaded, you may see volts of ripple, making it easy to assess the condition of the caps, or compare different ones.

Ed


Michael W. Lynch
 

On Fri, Feb 12, 2021 at 12:10 AM, Ed Breya wrote:


When it's sufficiently loaded, you may see volts of ripple, making it easy to
assess the condition of the caps, or compare different ones.
Being a novice at this game, this is why I asked. It may be anecdotal, but the original caps had much more ripple (unloaded) than was measured after the cap replacement by a factor of 10-20:1. My first indication of trouble was the malfunction of the PG506 and this has also been remedied. That being said, It is clear that I need to go back and measure the loaded ripple. I will do so and report the results.

--
Michael Lynch
Dardanelle, AR


Michael W. Lynch
 

On Thu, Feb 11, 2021 at 11:47 PM, Ed Breya wrote:


So, don't bother with no-load ripple - put a bunch of plug-ins in there to get
a better idea of what to expect.
Ed,

I did as you suggested and loaded up the TM506 with 5 plugins. I chose plug ins that utilized both the 11.5 and + or -33.5 rails. One slot reserved for measurement, of course. 11.5V (shared) is 680mV and +/- 33.5V are 340mV - 310mV respectively. There is no simple way to test the 11.5V on the HV compartment loaded, but that one would likely show improvement as well. So the 33.5 V rails now have less ripple loaded than they did previously unloaded and the 11.5V rail shows just about the same ripple, fully loaded than it did in the previously unloaded condition. So the ripple changed substantially with load, as Ed predicted. The one thing that I did not do previously was measure the ripple with the TM506 fully loaded and with the original caps. I think that would have been an interesting measurement.

--
Michael Lynch
Dardanelle, AR