Dana 5403 digital voltmeter repair

[valley001]

I have a bizarre interest in early digital voltmeters, so when I saw this thing rotting away on the internet I decided it must be given a loving home.

The front of this particular unit displayed an ominous "Do Not Use" sticker (orange for emphasis), as well as a missing line fuse and cap.   Not good signs, but surely its just a bit flaky. 



Borrowed a fuse and fuse cap from a parts unit and risked firing it it up.  Nothing.  Borrowed fuse blown.

Darn. 

Started working back from the PSU to the transformer.  Filter caps test OK.  Nothing seems out of place on the power supply.  I ohmed out the transformer, seems ok.  Decided to risk another fuse and tried powering the transformer alone...this time with a killawatt installed on the line cord.  Turned it on, amps rose slowly up to .5 and blew the fuse.  Darn.  Faulty power transformer I guess, but none of the windings were dead short.  Obviously something is awry.  Luckily I have a spare transformer in the parts pile.  Stay tuned.

[valley001]

Retrieved the spare power transformer...hmmm...its almost the same part number.  There is an "E" suffix added to the later 5403 model power transformer.  Otherwise it seems identical.  Primary ohms the same.  Secondaries, I have nothing to compare to.  The date code on the 5403 components place it 1969/1970ish manufacture while the 5500 model (parts unit) is 1967/1968.  My assumption (which I believe has turned out to be correct) is that these transformers are basically the same.  Innards certainly look healthier compared to the fried example. 



Replacement transformer fitted, tested PSU voltages...seem OK. 



Replacing the power transformer has coaxed some life out of the thing.  Swapped out some severely worn nixie tubes (see pics).  One of the digits was hardly visible due to decades of use.  Apart from seeing numbers and some basic functionality (like the decimal moving when changing range) it is not healthy.  DC will read negative values but put positive values it will zero out and stay on zero.  Ohms does nothing.  MV does same as DC.  Still much to be sorted.







It is drawing about 55 watts current, which is about 5 watts high according to the 5500 manual (max draw 50 watts).  The 5500 also has six nixies compared to 5 on the 5403, so I am probably drawing too much current.  Nothing seems to be getting overly warm but I think I may probe around more.  All of the PSU semiconductors and capacitors seemed to check out (but I completely missed the mark on the blown power transformer so who knows). 

Speaking of manuals, I have two versions of the 5500 manual (1968 and 1972 revision) but none for the 5403 and nothing turns up on the web.  Anyone have a lead on the 5403 manual? 

[valley001]

Much has improved.  Getting good readings out of the DC, MV, and Ohms functions.  Have not tested AC yet.  Re seating all of the relays did the trick. 

However:

A problem I now have is the replaced power transformer is getting quite warm after a period of time, almost hot to the touch.  The 5500 I have, which is a similar meter and draws close to the same wattage (50 watts or so), its power transformer does not heat up.   I am interested in any ideas based on what I have tried below.  The schematic is of the 5500 power supply but the 5403 seems mostly the same aside from missing the components for the 40v rail (no 40v rail on the 5403). 



1- I isolated the PSU and the power draw goes from 54 watts down to 5 watts.  I assume this eliminates the PSU as being an issue. 

2- R2 200 ohm resistor gets QUITE hot, and has evidence of heating.  It tests 200 ohms. .

3- Some of the decade boards were looking like they got hot in the past, I replaced them all with decades out of the 5500, and power consumption dropped by 2-3 watts or so.  It is still heating up.  What I assume are current limiting resistors on the display board are getting hot to the touch also..but perhaps this is normal? 

Voltages-

Pin 1: Ground
Pin 2: -17.8v
Pin 3: -34.6v
Pin 4: +29.4v
Pin 5: +17.5v
Pin 6:  No reading.  +40v rail, components not installed on this PSU (no 40v rail).  Presumably the +40v winding (White-White Brown) is present on the transformer.
Pin 7: +15v
Pin 8: +290v
Pin 9: +18.5v
Pin 12: +2.23v




Ideas? 

[valley001]

Hmmm....

Removed the decimal/function indicator board and power consumption drops by near 30 watts. 

[Kleinstein]

Too much current drawn by one of the circuit parts is the likely cause.
In theory there could be a broken diode or missing connection (e.g. one side of a split transformer), so that one rectifier is running only one way rectification and this way is less efficient, causing more power loss in the transformer for the same DC current.

The power for the decimal/function indicator board looks quite high - so chances are this could be the problem.

[valley001]

I do not have schematics for this board unfortunately.  What I think are current limiting resistors for the neons (130K big, carbon 2 watt maybe) are very hot to the touch, and the board underneath is burned to varying degrees.  There is one 250uf  electrolytic cap on this board associated with a network of diodes that testes 600pf, so I replaced it.  No change in current draw.  Current draw dropped by about 2 watts. 

With that board removed I see a 12 watt drop in over all power consumption however with that board removed the nixies each have two digits illuminated simultaneously.  My original estimate is high, I think, maybe this board is drawing around 17 watts assuming  about 1 watt per digit x5 nixies.  17 watts is still too high no? 

Too much current drawn by one of the circuit parts is the likely cause.
In theory there could be a broken diode or missing connection (e.g. one side of a split transformer), so that one rectifier is running only one way rectification and this way is less efficient, causing more power loss in the transformer for the same DC current.

The power for the decimal/function indicator board looks quite high - so chances are this could be the problem.

[valley001]

Going back through my work with replacing the transformer, I remembered the replacement is close to the same part number.  The original is 300017E and the replacement is 300017, and there is a physical difference.  The replacement has two sets of leads on the primary side, a set of two white wires and a set of two black wires.  (This is shown in the schematic posted above) They both measure the same resistance.  The original only had one set. 

I decided to join these leads (black with black, white with white) and connect the same way the original transformer was connected to the power switch.  With the leads connected I get about 12-13 ohms across the primary (measured from the wires connected to the line cord plug).  This is the same as the original.

While the unit was running and I was checking the heat from the transformer I noticed one set of these leads on the primary side of the transformer was warm/hot while the other was not. 

Could this be related to a load issue from the secondary or do I have other problems inside the transformer?   

EDIT-  I nixed the 40v rail on the psu from the 5500 model and installed it to eliminate PSU problems as a possibility.  Now waiting to see if it heats up. 

[valley001]

Hmm..reducing load by 15 watts still results in a very warm transformer.   

This unit, running at 50 watts- very warm transformer (44.7 degrees Celsius).
Another similar unit with the same transformer running at 60 watts- transformer not warm at all

Swapping PSU had no effect.  Voltage rails seem OK.  Unit seems to function otherwise. 

Maybe faulty filter cap?  I guess I need to look at voltage rails with my scope next. 

I am open to ideas here.  I did not think I would end up troubleshooting this sort of issue.

[valley001]

Here is the 300v rail wave form.  The specification is 500mv ripple, this is 10 volts.  The other rails are not as bad but some are a bit out of specification, this one is the worst offender.   C7 tests 70uf .8 ohms esr. 

[Kleinstein]

If using the transformer at 110 V, one may have to use both halfs of the primary in parallel. 50 W would need near 0.5 A and this still some 6 V drop for 12 Ohms DC resistance and this 3 W of heat from the ohmic resistance alone. Using only 1 half would double the loss.

The high ripple at the 300 V rail could indicate too much current drawn here, if the capacitor is still OK.
At least the diodes seem to all work (though still possibly too much leakage), as the ripple is 120 Hz with little 60 Hz.

45 C for the transformer may not be that hot.

[valley001]

Thank you all for the replies this far. 

I have focused attention back to the decimal/function board, thinking (as suggested above) that there is a current draw problem.  IIRC the 300v rail is taking 12 watts or so by itself which seems way too much for 5 nixies and 2 neons.    I took a proper heat reading from the large resistors on this board, as you can see they are topping out around 75 Celsius.  The board is burned underneath and there is evidence of heating on components surrounding the one very burned section.  Most all of these components seem to test fine, transistors, resistors etc.  Onc of the 130K resistors test 109K and one of the 2.2M resisitors tests 1.9M, but could this cause so much drama? 

Sadly, the DANA 5400 manual is close, but not a match for the 5403.  It is a match in almost every way except for the decimal/function board.  As you can see in the pics, the decimal board on my unit has a network of transistors and diodes associated with the resistors.  On the 5400 there are no such components, or they are located elsewhere.  I assume these serve as switches for the neons. 

5400 decimal/function board schematic:



5403 decimal/function board:



Burned section of the board.  Note the melted standoff on the nearby transistor:



Taking a temp reading on the 130K resistors (these are labeled 100K on the 5400 schematic).

[valley001]

This is my first attempt at drawing a schematic from a PC board.  Im not even sure what I have drawn makes sense, but this is my best guess so far as to how the transistors, resistors, diodes, and neons interact on the decimal/function indicator board (shown above).

Im am looking for insight into what sort of failure could cause these resistors (130K) to heat up so much. 

[Kleinstein]

If the schematics is correct, it is quite normal for the 130 K resistors to get quite hot. when off the resistors see nearly the full 300 V (that may be a little higher than nominal) and when the neon is on the voltage would still be at around 200 V.
If form some reason the resistor value is low this would cause even more heating. Carbon resistors could also have an even lower resistance when hot.
The schematics makes some sense, though not for making the circuit low power.

If the neons are well bright enough one could probably increase the values of the 130 K resistors a little.

From the pictures it looks like the last digit might be quite high intensity (could be an artifact from the photo).

From the ripple waveform one has a slope of about 1 V/ms for the discharge phase. With 70 µF this would mean a current consumption of some 70 mA  for the 300 V rail.

[valley001]

I think I made some discoveries tonight. 

I pulled out the DANA 5500 model to experiment with and learned:

1- The resistors on its display board also heat up to over 75C.

2- The transformer will heat up if I block the cooling fan inlet (the model I am trying to fix does not have a cooling fan). Got up to 40C before I removed the block from the inlet, it then cooled back to 33/34C.   

Back to the 5403:

I note evidence of heating of the front display window, it is warped and even appears almost melted in one place nearest the 130K resistors.  It certainly appears it was operated in an over heated condition for some time.

From this I am left to conclude:

The 5403 (a model that came later than the 5500) was manufactured without a cooling fan (cost savings maybe??) despite being similar in design to the 5500.  I believe the lack of cooling fan allowed the unit to operate in an overheated condition, which eventually lead to the failure of the primary coil in the transformer.

Agree disagree?

I am thinking of testing my theory and installing the spare 5500 back plate (with cooling fan) onto the 5403 and monitoring the transformer. 

In any case I suppose I will pause my order for filter caps from Mouser...

[Kleinstein]

Having a cooling fan can make quite some difference. The circuit looks like it uses a lot of power by design. So maybe running quite hot is normal and maybe more like a design fault, leading to "early" failure.  It would not be the first old days instrument that runs a little too hot.

Using a fan could be an acceptable fix.

I would consider replacing the 130 K resistor that reads low and maybe also the 3 others of similar type.  It does not have to be 130 K,  150 K should be OK too and run a little less hot.

[valley001]

 If I install 150K resistors where does the power go to instead of heat?   

Having a cooling fan can make quite some difference. The circuit looks like it uses a lot of power by design. So maybe running quite hot is normal and maybe more like a design fault, leading to "early" failure.  It would not be the first old days instrument that runs a little too hot.

Using a fan could be an acceptable fix.

I would consider replacing the 130 K resistor that reads low and maybe also the 3 others of similar type.  It does not have to be 130 K,  150 K should be OK too and run a little less hot.

[Kleinstein]

With 150 K resistors, there would be slightly less power drawn from the grid and thus likely also a little less loss in the transformer.
The slightly lower current would make the nones a little lower brightness. The difference is not great 130/150 gives just some 14%, or some 30% less than the resistor that now only reads 109 K.
150 K resistors may be easier to get than 130 K too.

[valley001]

I will give it a try.

I am wondering if I can do the same with the 20k and 39k resistors on the decade board.  R14-15 and R19-20 upper right corner of schematic below.