Hantarex MTC9000A resistor burning out

[bogbean]

I have a Hantarex MTC9000A chassis in an arcade machine I'm helping a friend with, we have R187 burning out, a 270k 1/2W resistor between B+ and the low end of the EHT coil. The resistor smokes almost immediately upon switch-on and burns out within about 20 seconds, so it must have a fair voltage across it.

I've attached a schematic and a section of it that I've annotated with some voltage measurements.

Is it fair to say the -1100V on the flyback side of R187 is far from correct? I'm thinking that the only way to get -1100V here is if there is a flyback problem on the EHT coil. I thought it shows the flyback must be being driven (or these voltages couldn't be generated). I wondered if there is a short somewhere in the EHT coil it may pull that point to ground and cause the pin 7 end of the coil to go very negative. B+ is badly affected but I think that's a symptom not the cause.

I thought I'd ask for some opinions before just blindly getting a new flyback to try. In-circuit resistance from flyback pin 7 to ground is only about 500 k.

The flyback is a HR Diemen HR7508, not the original for this chassis.

[agardiner]

Hi,

So I tend to agree that it looks like the LOPT has shorted out. However, I would personally want to carry out a few more tests first. I would disconnect pin 1 (drive) of the LOPT and hang a dummy load on there, say a 60W lamp. I would then want to scope the output of the BU508A, and also the input waveform, to ensure that the transistor is being driven correctly. It would also allow you to check the actual B+ voltage.

With the dummy load, don't run for too long, as the loading is obviously not correct for the output transistor. Pay particular attention to the drive waveform to ensure that proper cutoff is occurring. If these tests are all OK, then the LOPT is primary suspect.

[bogbean]

Hi,

Thanks for this suggestion. To check I understand I've made a sketch showing a bulb in place and LOPT pin 1 disconnected, is that the suggested arrangement?

With pin 1 desoldered and cleaned I can wedge it into the centre of its through hole so that it (just) doesn't contact the pad.

[agardiner]

Hi,

So yes, that is what I meant, but I have since rethought my advice, and think the following might be more useful, as the dummy load could rise to a false reading due to the lack of inductance. I think the following would be more useful....

1) Disconnect pin 1 of LOPT. Hang the dummy load across B+; ie, B+ to ground. This will hopefully (assuming no micro controller to shut the set down too quickly) allow you to check the B+ reading.
2) With the arrangement above, then scope the waveform on the base of the BU508A. As before this will allow you to check that it is a good clean waveform with proper cutoff.

Assuming both of these check out, the LOPT is still the primary suspect. Doing it this way though removes the possibility of blowing the BU508A due to the uneven nature of the replacement load, giving you more time to make the measurements.

[bogbean]

Hi,
I have tested with a 60W bulb across B+ and LOPT Pin 1 disconnected.
The bulb lit up (not very bright) and B+ was 123 V, not ideal but I took this as confirming the regulation circuit is working reasonably well.

I've attached a trace from the BU508A base. I took it to show that the transistor is being driven and does cut off but is not being driven to enough voltage to switch the transistor on. I'm wondering now if my probe had slipped to the 1x position.

I was puzzled as to how the TDA2595 is powered without the LOPT running so I went to measure the 12V rail but mistakenly connected B+ there instead of my DMM There's a 12V zener in this rail that has now gone short, I'm hoping that may have limited the damage, I need to investigate.

Whilst having a good visual inspection I can see cracks in the LOPT so I do think that will need replacing. I just hope i can fix the damage I caused with B+!

[agardiner]

Hi,

OK, so 123V B+ is not a deal breaker at this stage as the lamp is of course an artificial load. Fingers crossed with the 12V rail; hopefully just the zenor!

There is enough voltage driving the BU508A and the frequency is correct. Would nice to check that it is ramping correctly, so also check the waveform at the driver transformer and check D110. However it is looking very likely that the LOPT is the culprit and your visual checks also point to this. Once you have fixed the 12V rail, I would be changing the LOPT and powering up with a limiter.

[agardiner]

Although limiters can often cause the power supply to blow up as they try to regulate, so probably better to just power up once items replaced.

[Chris55000]

The nominal h.t. is +130 V marked on the MTC9000/A Circuit, so your reading of +123 V will cause lower e.h.t., rather than a higher one!

Make sure the BYV95C beam current limiting diode D117 and capacitor C152 are not open circuit, and if these test OK, there is almost certainly a defective diode or e.h.t. winding in the LOPT and you'll need to get another one! It's more usual for e.h.t. section failures to burn up the 1k/2k2 directly connected to the bottom end of the e.h.t. winding rather than the 270K bleed resistance down from the h.t.+ line tho!

The one I can find listed for the MTC9000/MTC9110 is type 1105.E048 or HR Diemen HR7107, but the type used in the '9000A isn't marked on the circuit, I assume it was previously working with the HR7508 – HR Diemen list both '7107 and '7508 used in this chassis!

Chris Williams

PS!

I am aware the MTC9000 and MTC9000/A are different – I did ten years myself repairing PC monitors!

[bogbean]

The zener arrived today so I will have another go at the weekend.
D117 reads ok and I have previously checked C152 as ok out of circuit.
R186 (2k2) also reads ok.
Yes the chassis was previously working with a HR7508 although I did see on the datasheet that the 7508 generates some different voltages to those shown on the MTC9000A schematic.

This is the first chassis I've ever worked on, I have been mulling over this problem for a while now and previously suspected the lopt. I did a ring test back in April and I felt it looked ok because I counted at least 16 oscillations. Am I right that a ring test can detect shorted turns but perhaps not all failures?

[Chris55000]

Yea – a "ring test" wouldn't show up an open circuit diode in the e.h.t. section, or a short between successive sections of the e.h.t. winding if it only shorted one or two turns out, nor would it detect insulation failures!

Many places, including Dowding and Mills where I repaired Monitors until I got made redundant by the closure of the electronics department there, used to keep the "HR Diemen Flyback tester" that generated half–sinusoid positive pulses with a p.r.f. of about 30 kHz ("monitor" version of the tester) which you connected across the primary, insert the anode cap in the clip and connect the earth lead to the bottom end of the e.h.t winding – HR Diemen's database shows where, the voltmeter whose indication is sbown on the LCD is calibrated specially to allow for the much lower drive pulse amplitude – but unless you're still regularly doing arcade games, etc., it's not worth buying one for just one transformer – you might as well buy another transformer!

Glossary

p.r.f. = pulse recurrence frequency

Chris Williams

[bogbean]

Thanks Chris, I'm a newbie so it's good to understand the limitations of various methods.

We'll get another lopt, Donberg and ebay seem to be options so will make a decision between those two.

[bogbean]

I've fitted a new LOPT and the monitor is working now . I'm thrilled and also relieved following my 16th May voltage error!

B+ measured at 125.3V, which I think is ok (based on the 130V +/-3% tolerance given in the MTC9000 service manual, I don't have a 9000A service manual).

Many thanks for your help Chris and Adrian.

Prior to the LOPT problem I had to replace TR2 and TR9 (we had lost red and blue colours and both these transistors had high Vbe) on the neckboard so I think it would be good to check the +24V voltage rail to see if there was a common cause for these transistors failing.

[agardiner]

Excellent. Glad you have got it working again!