7904A compress sweep


Tom Phillips
 

I need some advice on where to start looking to track down a compressed sweep issue on a 7904A I was given.

The sweep is compressed by about one division on each end and it does not matter what plugins or channel combinations are used. The readout seems slightly compressed as well.

Very rarely it will suddenly appear normal width and readout seems to slightly expand but then seems to revert shortly after.

Its a very nice scope in otherwise excellent condition.


Harvey White
 

I would guess that the horizontal drivers are bottoming (or topping) out.  This suggests a power supply voltage is off.  That it suddenly goes "good" then back again suggests that the *effective* power supply to the horizontal (or vertical, too) amp is changing.

I'd measure the power supply voltages both on the deflection driver boards and the supply itself.  Problems could be bad connections, a filter capacitor acting up, an odd filter resistor (if any).

It could be just as easy as a transistor (in the input circuitry) with an intermittent contact.  I'd be checking the static voltages with the intensity turned down to zero and no sweep.

Harvey

On 7/16/2021 3:53 PM, Tom Phillips wrote:
I need some advice on where to start looking to track down a compressed sweep issue on a 7904A I was given.

The sweep is compressed by about one division on each end and it does not matter what plugins or channel combinations are used. The readout seems slightly compressed as well.

Very rarely it will suddenly appear normal width and readout seems to slightly expand but then seems to revert shortly after.

Its a very nice scope in otherwise excellent condition.






Precaud
 

Sounds like it could be as simple as an intermittent pot in the horizontal section.


Harvey White
 

Good thought, could be that, but if the gain setting pot was intermittent, then I would expect it not to clip, since the amplifiers are designed to enable off screen deflection.

Perhaps another one?

Harvey

On 7/17/2021 8:14 AM, Precaud wrote:
Sounds like it could be as simple as an intermittent pot in the horizontal section.





Tom Phillips
 

Thanks for suggestions

I moved the gain pot slightly but it has little effect.

The power supply voltages are very close to spot on and stable but the four metal output transistors Q610, 20 30 and Q640 are definitely hot. I'm in the middle of a long distance move and don't have another scope handy at the moment to check stage waveforms but if needed are equivalents for motorola ST919/ST7394 available? Casual web search not turning up much.


Harvey White
 

If the trace can center, then the circuit is in balance as far as the + section and - section are concerned, putting the same voltage on the plates.

There is, however, no guarantee that it's the *right* voltage.

I'd center the trace, or duplicate the test conditions in the manual, then go through and check the static voltages on each transistor.  Running each output at, say, twice the current will still give you deflection, but it will bottom out, for instance.

Did you check the emitter supply resistor (and bypass capacitor) for the emitter supply circuits of the balanced pairs?

Harvey

On 7/19/2021 8:46 AM, Tom Phillips wrote:
Thanks for suggestions

I moved the gain pot slightly but it has little effect.

The power supply voltages are very close to spot on and stable but the four metal output transistors Q610, 20 30 and Q640 are definitely hot. I'm in the middle of a long distance move and don't have another scope handy at the moment to check stage waveforms but if needed are equivalents for motorola ST919/ST7394 available? Casual web search not turning up much.







Mark Vincent
 

Tom,

Use 91% IPA for the coax inputs and socketed transistors. I use a small artist brush to apply the IPA then move then move them up and down in their sockets a few times to clean them. You do this to the neck pins and sockets also. Dirty contact(s) would be the most likely cause of your trouble.

I think you mean the hot transistors are the 900 numbers. I have three two fin Thermalloy 2225C heatsinks on three of them. There is not enough room for all four to have this type of heatsink on them. The heatsinks on mine are on Q900, Q930 and Q940. I used MX-4 compound.

In mine, I replaced C100 with a film type and the two 22mfd with ULD 47mfd 25V. R921 and R954 are now, in mine, 1W 1% and R900, R941 and R960 are 2w 1% types. I had these parts in stock. These resistors are off the board to allow air all around the bodies of the resistors. I did add compound on the two pars if transistors with the metal band around them to further aid in thermal coupling. You do what you want with your piece. I did what I wanted to in my piece. Mine works fine.

My B time base socket on the main board had bad/no solder on the two coax output from the time base. This was a factory defect. No one had worked on this after it was originally bought. The spring tension wire was off its pin. I took these off, repaired the problems then put back in. Now it works. The two plastic covers were carefully removed to do the work. I know what I had is a rare defect.

Mark


Tom Phillips
 

Solved!!

Thanks for all for the suggestions. It help me puzzle this out.

The issue is a wacky 500 ohm gain pot R230.
Measuring it I get erratic values above 500 ohms so I tapped the adjustment screw and bingo 195 ohms. That was causing the emitter bias for the differential pair to be out of range. Before I replace it Ill reflow the joints in that area and see if it fixes it.

Also I'm taking Marks suggestion and adding heat sinks to the outputs. They are just too hot for comfort even with the fan.


 

On Mon, Jul 19, 2021 at 05:49 PM, Mark Vincent wrote:


think you mean the hot transistors are the 900 numbers. I have three two fin
Thermalloy 2225C heatsinks on three of them. There is not enough room for all
four to have this type of heatsink on them. The heatsinks on mine are on Q900,
Q930 and Q940. I used MX-4 compound.
My 7904A service manual does not contain a Q900. The output transistors are Q910, Q920, Q930 and Q940.

The 'scope - and many of its siblings - has lived and survived for many decades. Feels like leaving well enough alone to me, especially giving following considerations.

Thermalloy 2225C fins are not well suited in these situations: Their thermal resistance to the environment is not very low and they need a lot of space.
The star-like variety with radial fins would be much better. These fins could interleave on the PCB. Tek has often done that. Not critical because each set (Q910/920 and Q930/940) has collectors connected together.

I wouldn't like three out of four finned, leaving the fourth bare (and hotter). Not sure if it makes a lot of difference but I'd prefer going symmetrical, like the amp's config, i.e. either Q920/930 or Q910/940, if only two fins are available.

The fins are connected to the collectors (case) and add some (parasitic) capacitive coupling to the environment. That could well influence frequency characteristics of the amp, especially noticeable at high horizontal speeds.

Raymond


Mark Vincent
 

Tom,

If your pot is bad. I do have some 7mm 500 ohm pots that should fit. If you want one, contact me and I will send you one.

Mark