[FIXED] Pintek Oscilloscope sweep slows down at slow timebases

[serafis]

Hi, hoping someone can help. I have acquired a Pintek PS-200 20MHZ oscilloscope and am attempting to calibrate it. So far, everything has been going fine - PSU voltages check out, brightness & focus fine, DC Step balance, DC Balance and 5x magnification, B-channel invert and Add(A+B) DC-gain all set Ok. Input capacitances and phase compensation all set, 100MHZ square wave all set Ok too. Speeds faster than around 10ms are all OK too. Square waves and sine waves above this speed look good, with correct vertical and horizontal displacement. So far so good.

But for speeds slower than 10ms , the trace (dot) slows down as it moves across the screen, and at the 2-second speed it actually stops at around 7/8ths of the screen and sometimes even moves back very slightly.

The sawtooth ramp off the sweep generator looks straight at speeds faster than 10ms, but curves off at slower speeds as per the attached image; the curve becoming flatter the slower the speed.



I've removed the sweep generator board and checked all diodes, resistances and capacitances, replacing a couple of electrolytics with no effect. I could really use help from someone more expert than I as to how to diagnose the issue further and what to check for,

Schematic of the sweep generator also attached.

 11_b_PS 20-60mhz_compressed.pdf (26.55 kB - downloaded 42 times.)

Thanks,

Jon.

[bdunham7]

Did you replace C710?

[serafis]

Thanks, yes, that was my first thought - I replaced C710, C709 and just to be sure, C721 as well. No difference at all unfortunately, and the removed capacitors tested OK anyway. Just a point, the +12V rail is exactly 12V, the -12V is 11.87V so very slightly down. Both of these come straight off 12V regulators after rectification and smoothing in the LV power supply. I was wondering about replacing the -12V regulator just in case, but since all other functions and measurements appear spot-on, I couldn't see how this would affect only a slow sweep.

[bdunham7]

Actually, now that I look at again, C710 is probably not the one I meant.  Right above the 'TIME C' label, there are two 47uF caps, one looks like it says C710 but it probably isn't.  Perhaps both of those deserve a look.

[serafis]

Ah Ok. Interesting, neither of those caps (C718 / C718A) looks like an electrolytic - they're yellow 'beads' (tantalum?) as per the attached photos. And they're marked 33 | / 16+ so presumably 33uF rather than 47uF, and not sure if they're polarized.  Also C718B, which isn't shown on the schematic, but is in parallel with the other two on the board, is the same. So in the schematic, 47uF+47uF = 94uF, on the board 33uF+33uF+33uF = 99uF. so perhaps the board is a slightly modified design. Anyway, close enough do you think? Or should I replace C718/718A with 47uF electrolytics and remove C718B?

[bdunham7]

If I were going to test those by substitution, I'd just remove all 3 and replace one with a 100uF electrolytic.  Note that they are polarized, but they mark the + side, not the -. 

However, I had a closer look at your schematic and I see by the timing diagram in the upper left that for 10ms, C715 is the timing capacitor.  The big 100uF section doesn't come into play until 100ms/div.  If your sweep is actually good for all speeds down to 5ms/div and then is bad for all speeds 10ms/div to 2s/div, then there's no single component that falls out when you look at that diagram.  So perhaps go back with your other scope and look at the sweep pattern for everything from 500us/div to 1s/div.  If it is exactly as you have described, then I'm a bit stumped for now...

[serafis]

I measured the 3 capacitors and they all appear to be ok. I've reassembled the scope, and using my other scope looked at the sawtooth waveforms coming off the sweep generator board for each time division, displaying one complete cycle on the screen for each division. I can see no perceptible curve from 0.1uS to 2mS. then perhaps a tiny bit at 5mS and then increasing with each time division to 2S, where the sawtooth actually decreases very slightly from its peak and never ends. The trace of course reflects this, with movement across the screen becoming slower until 2S when the dot stops somewhere near the middle of the screen (and sometimes moves slightly back).

The peak-to-peak voltage of the sawtooth starts at around 6.3V at 0.1uS, decreases slightly to 6.1V at 1uS, remains around this level until 0.5mS then drops to around 5.7-5.8V at 1mS and remains around this level through the rest of the time divisions, although the increasing curvature of the ramp means it takes longer and longer to reach this level as the ramp becomes less linear, and eventually flat at 2S. The supply to the sweep generator remains constant at +12V/-11.86V  through all time divisions.

I'm a bit stumped too. I'm wondering if it's something like a "leaky" diode or transistor, where the leak isn't noticeable at faster speeds; or a capacitor that is losing a bit of charge so can't fully charge in time at these  slow speeds,  but I have no idea where to look for this or how to find it. I'll try swapping out the three capacitors for a single 100uF 16V one, to see if it makes any difference at all, but I'm not too hopeful.

[bdunham7]

It might be worth looking at what the +12 and -12V supplies are doing near Q704 and C723.

[serafis]

Well whadda you know! Voltages on Q704 and C723 were fine. Then I checked the waveform on the +ve leg of C718 and it was the curved sawtooth. Aha I thought, maybe they're not charging properly. So I pulled out those three caps and checked them on my capacitance and ESR meter and they were way, way off, although they'd tested ok in circuit. One was just about OK at 32uF and 0.4 ohms ESR, the second was around 40uF and 6.5 ohms ESR and the third was 52uF and 7.5 ohms ESR. So I replaced them with two good quality 47uF 16V electrolytics as per the schematic, put it all back together and presto! a beautiful sawtooth ramp at 2S as per the attached pic, with the dot moving at a constant speed (at around 2S per division) right across the screen. So all fixed now.

Now all I need to do is reassemble properly and recalibrate from scratch, which takes a while but isn't too difficult, and it should be perfect. Thank you so much for your help, couldn't have done it without you!