Another Tale of Two Hameg 20MHz scopes.

[Zenith]

I mentioned these before but not in any detail. The 203-6 came from a swapmeet for £15 but it did include a pair of HP 10071A probes in a nice condition, which are worth at least that much. They are some of my favourite probes. They are robust without being chunky. The 203-5 was bought with a Hameg 8028 spectrum analyser which came with the special BNC cable, which is as rare as hen's teeth. The seller reckoned that the scope and the spectrum analyser were matched in some way, so had to be sold as a pair, which is b/s. I freed up the controls and lubricated them, cleaned switches, checked the low voltages. I changed the problem reservoir capacitors in the 203-6. I checked all functions and adjusted the channel gain and they were in basically working condition. I dug them out again because I've promised to give one away and the time is approaching. It's nice to give friends things that work. Hameg never produced a detailed factory set up procedure for most of their scopes. They give a few hints as to what the presets do and a sequence for common adjustments.

Hameg 203-5.

All low voltages come from 78XX regulators. They were spot on with no excessive ripple. The 170V line was adjusted to be 170V. This left the -1900V line, which is adjustable. I tried measuring it with an AVO which goes up to 3kV and it showed 600V and wandered a lot. Another AVO which goes up to 2.5kV produced similar crazy results. They probably both have faulty high voltage resistors. I searched the box of high voltage resistors and found one which measured 45.93M. My oldest DMM has a resistance of pretty much spot on 10M on all voltage ranges. 10/55.93 = 0.179. Putting the resistor in series with the meter should give a reading of 0.179 x the actual voltage. It worked with a bench power supply. The EHT of the scope had to be adjusted to show 339V on the meter with this arrangement, with hands off technique of course. The EHT was a bit low and was adjusted to show -339V on the meter. This kind of arrangement would be a pain for regular use, but it's OK for once in a while measurements. There's a granularity error of about 5V at 1900V. I guess that the tolerance of the -1900V line is not better than +/- 1%. It was about -1815V before adjustment.

The other thing Hameg recommend is checking the blanking pulse, which is 22V ptp superimposed on -1900V. They give waveforms and say a clean square wave should be aimed for. They suggest a 10nF 2kV capacitor between the test point and a 10X probe. I don't have a proper HV probe. I found a 4.7nF 4kV ceramic in the box of HV capacitors and used that. I set the preset somewhere in the middle of the range where it produced the clean square wave.

IMG_2282.JPG

I went through the set up, paying more attention to getting a clean 1MHz square wave and making sure X-Y mode, triggering, the component tester, and calibrator etc, were set properly. The next day I turned it on for a final check and found the display was doing strange things. Sure enough there was heavy ripple on the +12V line. I changed the two 1000µF 25V caps I really should have changed before, and a few other PSU caps, and the problem went away. Some of those caps are hard to replace without removing a PCB, which is a serious mission. I checked the set up against a Siglent AWG and it was pretty much spot on. The scope's as good as I can get it and I'll give it away with a reasonable expectation it won't act up.


Hameg 203-6.

I'd changed the 1000µF reservoir capacitors on this. I didn't do more than already stated by way of setting it up and checking it. The EHT and blanking are not adjustable on this model. The PSU changed from 78XX regulators in the 203.5 to an L200 for the 12V supply, The 150V and -12V supplies are discrete and are referenced to the +12V line. There are a couple of different 5V supplies, which are discrete and produce low current. One for logic chips and the power indicator LED and the other for the Y preamps. They are basically emitter followers powered by the +12V line. Only one is brought out to a header where the other voltages are presented. The manual doesn't give the position and use of the presets, but the PCBs are silk screened.

It worked but started to show a fuzzy display. There was 400mV ripple at 1.8MHz on the 12V line. I've seen this problem before. I changed the rest of the electrolytics in the +/- 12V PSU, which are 10µF 35V, except I used 47µF 25V after checking. Some of these are a pain to access and need keyhole soldering. The problem went away.

Then when checking the vertical amps I noticed there was erratic amplitude on channel 2. I prodded around with an adjustment tool and it was one of the ribbon connectors to the Channel 2 preamp. That was resoldered and that problem disappeared.

Then I set it up for the best square wave at 1MHz, which the silk screening hints should be done. Sporadically, the amplitude of the two traces would change, especially channel 2. I investigated the supply lines at the header and occasionally it would produce random dips of about half a volt, which I could catch on a DSO. I wasted some time looking at the wrong 5V supply, before I traced through and found the second one.

5V-supply.png

I remade all the soldered joints likely to cause a problem. No change. I changed the electrolytic. No change. I thought ceramic capacitors rarely fail, especially when there's 5V across them. Resistors can act up. I changed the two 4.75K resistors for two matched 1% 4.7K metal film types. No change. I changed the BC237B for a BC547B, which seems a slightly superior replacement. No change. I changed the two ceramic capacitors for two from the ceramic capacitor collection and the problem disappeared. It's the first time I've seen a problem like that with ceramic capacitors. They didn't seem physically damaged. They operate at low voltage.

Anyway, I started looking at the calibrator with a 10X probe. It has 2V and 0.2V outputs. If the probe was compensated on one it wasn't compensated on the other. It was a problem with the attenuator. I've always regarded attenuators as a black art and I've been happy to not have to mess with them. I found this https://entertaininghacks.wordpress.com ... -response/. A site of immense erudition and definitely the work of one of the foremost engineering minds of the modern age. It links to a discussion on Tekscopes.io which explains setting up attenuators. You can't do it with a sig gen with a 50 ohm termination. You need something which has the same capacitance as the most sensitive V/cm range on the scope - that usually has no attenuation. That's the basic input capacitance of the scope and the attenuators are adjusted to that. It should be the same for both channels and is usually close. I don't believe many readjust the probe compensation when moving a probe between channels. If you are using a range which puts two attenuator sections in series, incompatibilities cause large compensation problems. A 10X probe will do it, but you'd need a sig gen with a very high ptp signal for the highest V/cm ranges. The thing to use is an input normaliser, which is a 1M resistor with a trimcap in parallel. It's a 2:1 adjustable probe.

This one was made using a little aluminium box that was bought from AliExpress. Sometimes there are official Tek input normalisers on ebay for about £20.

DSCN3617.JPG

About a year back I bought a Waugh oscilloscope calibrator. This has various outputs such as for TV sync, but the two of immediate interest are the square wave generator which produces a 1kHz signal with convenient preset levels from 0.12mV to 120V ptp, and another square wave generator which produces a <1ns rise time square wave at up to 2.5V ptp and periods from 5s to 0.1µs. Usually the trace on the scope needed for adjustment is 6 divisions high. Using this and the normaliser makes it easy to adjust the attenuator for input capacitance and frequency compensation on all ranges. Checking with a 10X probe as best I could confirmed that the compensation no longer changed with the V/cm range, as it should be.

DSCN3888.JPG

Some scope manuals recommend the use of an LCR meter to set the input capacitances for the channels. I think setting the input capacitance using a signal generator and the input normaliser is easier and more certain.

I went back and checked the 203-5 attenuators. Only the very slightest tweak was needed on a couple of presets. Attenuators don't often need adjusting. I wonder if the ones on the 203-6 had been tinkered with. I redid the 1MHz square wave on the 203.5 using the oscilloscope calibrator. It's very hard to get the square wave right. If it's as close as it can be got at 6 divisions, it has horrible overshoot at 3 divisions, so a compromise is needed.

Both scopes showing a 1MHz square wave from the calibrator. All functions work properly and they are as good as I can get them. The 3dB down point on both is over 20MHz and they trigger to at least 40MHz.

203.5

IMG_2270.JPG

203.6

DSCN3870.JPG

The HM 203-6 was a long haul with with problems developing along the way. I learned something about attenuators. I've had enough of Hameg scopes for now. I'll concentrate on proper scopes from now on. I wish I'd turned to the Waugh calibrator before.

[tggzzz]

The other thing Hameg recommend is checking the blanking pulse, which is 22V ptp superimposed on -1900V. They give waveforms and say a clean square wave should be aimed for. They suggest a 10nF 2kV capacitor between the test point and a 10X probe. I don't have a proper HV probe. I found a 4.7nF 4kV ceramic in the box of HV capacitors and used that. I set the preset somewhere in the middle of the range where it produced the clean square wave.

I have a P6013 which you could borrow, if necessary. 12kV, 1000:1. It would be a pleasure to help someone with such impeccable taste

Before in had that I debugged a problem with a Tek465 blanking pulse. Basic technique was to hang an voltmeter off the grid and kathode, and another off the blanking pulse. The two voltages should be the same, and very in the same way as the timebase, holdoff, and delayed timebase highlighting vary. In !y case they did, because the fault was a broken connection inside the CRT.

I did not, of course, hang 3*33Mohm + 100kohm resistors off the 2450V grid, and measure the voltage across the 100kohm. Oh no.

[bd139]

Good to 2KV!

7kSoeG3.jpg

[tggzzz]

Good to 2KV!

Yeah 80% is good enough

[Specmaster]

Those little Hameg scopes are nice bits of kit in the bargain, too.