5440 PS transistor selection


Dave Peterson
 

Finished replacing the big electrolytic caps in a dead 5440 power supply yesterday.

In the interest of testing it before reassembling the whole scope I tried bench testing it with a variac. It was doing ok as I came up to 35v, but the negative 30 and 15v rails weren't coming up. At least not as much as the positive supplies which were about linearly related to input level. And something was putting out a hell of a lot of heat near the -15v regulator. Current was stabilizing at about 400mA.

I shut it down and using my calibrated finger found Q880, -15v output, was warm. The other output Qs were not. I brought it back up to 35v and finger tested the other Darlington output Qs which were all cool. Except for Q880, which was almost like touching a soldering iron.

Shutdown, pulled Q880, and checked on the curve tracer - definitely bad.

The Replaceable Electrical Parts list shows: Q820, Q850, Q880 as 151-0405-00, "sel from MJE800".

I see Walter at Sphere has a few, Mouser has 4300 on back order with a 43 week lead time (O_o). I suppose I can order the three from Walter, or keep searching. I'm sure I'll be able to find something. They don't appear prohibitively rare or expensive. And/or there are other equivalents.

The question is: what is meant by "sel from MJE800"? What am I selecting? Do I have to install, test, and adjust the -15v to +/- 10% by selection of MJE800s?

I'm sure that was fine for Tek technicians with a bin full of them at hand, but these don't seem plentiful online. Or is there something else I'm supposed to be selecting for?


ppppenguin
 

The MJE800 is a very ordinary NPN darlington: https://www.onsemi.com/pdf/datasheet/mje700-d.pdf
Any NPN darlington in a similar package rated at 60V and 4A or better should work. Using the Mouser search facility shows several candidates such as: https://www.mouser.co.uk/ProductDetail/ON-Semiconductor/BD677G?qs=vNc2DXHODiIKZu%2FUBUHHAw%3D%3D

As for selection, I suspect it was for minimum hfe. I wonder why this was so critical in a darlington which has a high hfe anyway. Could have been for voltage, in which case just spec an 80V or higher part.


Dave Peterson
 

Thanks for the cross reference. That helps.
My wondering about selection was more about, is the output level tuned to +/- 10% by trial-and-error?
Seems odd. If it were for min/max hfe, that'd make a little more sense, and something I can do with tools before wholesale reassembly. These three are not difficult to replace in a fully assembled scope. still, not thrilled with the trial-and-error methodology/prospect. Wouldn't that make it somewhat load dependent?
It's also interesting that these three Darlingtons are listed as "Sel", but the +/- 30v supplies, which are a different transistor type, and have a variable pot, don't specify "select". Because those outputs are adjustable?

On Friday, June 4, 2021, 07:25:13 AM PDT, ppppenguin <tekscopes@borinsky.co.uk> wrote:

The MJE800 is a very ordinary NPN darlington: https://www.onsemi.com/pdf/datasheet/mje700-d.pdf
Any NPN darlington in a similar package rated at 60V and 4A or better should work. Using the Mouser search facility shows several candidates such as: https://www.mouser.co.uk/ProductDetail/ON-Semiconductor/BD677G?qs=vNc2DXHODiIKZu%2FUBUHHAw%3D%3D

As for selection, I suspect it was for minimum hfe. I wonder why this was so critical in a darlington which has a high hfe anyway. Could have been for voltage, in which case just spec an 80V or higher part.


ppppenguin
 

I don't know the details of that circuit but unless somebody at Tek had been smoking something well out of the ordinary they certainly wouldn't fine tune the output voltage by selecting transistor parameters.

No sane designer would ever select for max hfe. In an ideal world you don't want any base current so a perfect transistor would have infinite hfe. Real life isn't so kind so you need a minimum hfe so it doesn't upset the circuit driving the base. With a darlington I can just about conceive a designer might need to select for min hfe but it seems unlikely.

My view is just replace with something that's similar and try it. If it works within tolerance and doesn't get horribly hot then you've fixed it. You say there's a +/-10% tolerance on the output voltage. That's big enough to drive a Sherman tank through. In any plausible design the voltage will be set by a reference (such as a zener or another stable voltage rail) and some resistors. If you can't hit +/-10% without trimming then that's like not being able to hit a barn door while standing inside the barn!


Dave Peterson
 

You put a grin on my face. :)

Yeah, there's that. Not uncommon for me to fret details when the right answer is: it's fine Dave. Swap the damn transistor and get on with your life! :)

Need to build another Mouser order. The shipping is more expensive than 10 transistors. :P

Thanks for your inputs!

On Friday, June 4, 2021, 07:57:06 AM PDT, ppppenguin <tekscopes@borinsky.co.uk> wrote:

I don't know the details of that circuit but unless somebody at Tek had been smoking something well out of the ordinary they certainly wouldn't fine tune the output voltage by selecting transistor parameters.

No sane designer would ever select for max hfe. In an ideal world you don't want any base current so a perfect transistor would have infinite hfe. Real life isn't so kind so you need a minimum hfe so it doesn't upset the circuit driving the base. With a darlington I can just about conceive a designer might need to select for min hfe but it seems unlikely.

My view is just replace with something that's similar and try it. If it works within tolerance and doesn't get horribly hot then you've fixed it. You say there's a +/-10% tolerance on the output voltage. That's big enough to drive a Sherman tank through. In any plausible design the voltage will be set by a  reference (such as a zener or another stable voltage rail) and some resistors. If you can't hit +/-10% without trimming then that's like not being able to hit a barn door while standing inside the barn!


 

Dave,

for clarity, when the Tek parts list says it is selected, it means that Tek found the generic specs too loose to insure correct operation, so they specified one or more parameters that had to be tested to insure that. In the case of the MJE800, it is almost certainly beta, as the spec is incredibly wide, and Tek will have wanted at least a threshold beta at a specific current. All darlingtons have some issues with this parameter, and this is why they are almost never used in audio circuits, the variation is just too wide.

the Tek 151-spec for the transistors says exactly what the selection criteria are, you can download the transistors specs from the Tekwiki to get an exact answer. What is more puzzling is that transistor is hot, unless there is an unexpected load on the downstream side, or collector short to chassis ground, this probably shouldn't happen. Not so easy to check a darlington with and ohmmeter, as it may have a CE shunt diode and bias resistors, which really complicate things.

Are you sure there is no issue AFTER the transistor? Worth it to really check, as bypass caps seems to be the eternal nemesis of all Tek hardware. If the part really is bad, let me know, I will drop one in an envelope to you for free.

all the best,
walter (walter2 -at- sphere.bc.ca)
sphere research corp.


Tim Phillips
 

I am following this thread with interest - I am back to working on one of
my 5441s and finding that the +30V is up at +38V (not regulating) and won't
adjust, and the -15V is down at -8V. Manual says -15V sits between +30V and
-20V.. I checked all the other Darlingtons and did a quick BCE / ECB check
in case of previous tinkering. Also checked the insulating pads between
Q850 / Q880 and chassis.
I think it was Dennis who remarked on the fallibility of some ageing
Zeners, and I also notice one of the 'fat' 0.2 Ohm resistors is high (0.4)
Yes, I am trying to talk myself out of replacing the bulk capacitors !!!
Tim

On Fri, 4 Jun 2021 at 23:24, walter shawlee <walter2@sphere.bc.ca> wrote:

Dave,

for clarity, when the Tek parts list says it is selected, it means that
Tek found the generic specs too loose to insure correct operation, so they
specified one or more parameters that had to be tested to insure that. In
the case of the MJE800, it is almost certainly beta, as the spec is
incredibly wide, and Tek will have wanted at least a threshold beta at a
specific current. All darlingtons have some issues with this parameter, and
this is why they are almost never used in audio circuits, the variation is
just too wide.

the Tek 151-spec for the transistors says exactly what the selection
criteria are, you can download the transistors specs from the Tekwiki to
get an exact answer. What is more puzzling is that transistor is hot,
unless there is an unexpected load on the downstream side, or collector
short to chassis ground, this probably shouldn't happen. Not so easy to
check a darlington with and ohmmeter, as it may have a CE shunt diode and
bias resistors, which really complicate things.

Are you sure there is no issue AFTER the transistor? Worth it to really
check, as bypass caps seems to be the eternal nemesis of all Tek hardware.
If the part really is bad, let me know, I will drop one in an envelope to
you for free.

all the best,
walter (walter2 -at- sphere.bc.ca)
sphere research corp.







Dave Peterson
 

Tim,

As unpleasant as it is, I'm becoming an electrolytic replacement advocate. Even today electrolytic capacitor lifetime is on the order of 10 years, and these 70's vintage scopes are obviously beyond that. Having said that, I'm also of the "if it ain't broke, don't fix it" camp.

There are a few things about the 5440 that had me abandon the electrolytics right off: their age of course, then there's something about these 5000-series scopes and TM500 mainframes that suggest a more budget minded design. Not throwing shade at Tek, I just wonder if the caps are of lesser quality, and if the system is more stressful on them. And finally, the rectifier for the +/- 30v rails is the ubiquitous 152-0488-00 - a known marginal original part that fails frequently in 465s. I just replaced one last night on a 465 that is in otherwise great condition. Now working beautifully again. I wonder what its failure mode does to C930/C932 (are they the same number for 5441?). Again, are these caps more stressed? The 152-0488-00 has a readily available modern replacement. Parts list coming.


So with all that doubt in my mind, especially with an already failing supply, to me, it's just not worth it to not recap. And in so doing I pulled the entire PS out of the body so the transformer, main board, and plug-in board are all still connected. It's a bit of a pain to handle, but it is vastly easier to work on without the frame rails in the way.

And a final point I want to share after having done it: cut up and destroy the old caps getting them out. After fighting with a soldering iron and doing a fair amount of board damage I gave up on them. The difficulty is getting them cut, but patience, care, and good tools does the job. Work from the outside in. I used a rotary cutter on a Dremel tool with a flexible extender. Carefully! I'd like to find some smaller hand saws maybe. Perhaps something from Exacto? I'll have to do some research. The Dremel was rather scary, but I got it done without hitting anything else.


I've done a couple of recaps now (still no expert), and these old cans just aren't worth saving. Not when the cost is the cumulative damage to the more valuable board. With the body gone the solder flows much more quickly and you stand a better chance of not pulling out the hole metal. I have some copper tape for trace repair, but it's much better to not damage the traces in the first place. Screw the old caps! Grin.

Recapping is a royal pain. But I personally feel that it's ultimately worth it if your PS is already dead.


Dave

On Saturday, June 5, 2021, 2:43:59 AM PDT, Tim Phillips <timexucl@gmail.com> wrote:

I am following this thread with interest - I am back to working on one of
my 5441s and finding that the +30V is up at +38V (not regulating) and won't
adjust, and the -15V is down at -8V. Manual says -15V sits between +30V and
-20V.. I checked  all the other Darlingtons and did a quick BCE / ECB check
in case of previous tinkering. Also checked the insulating pads between
Q850 / Q880 and chassis.
I think it was Dennis who remarked on the fallibility of some ageing
Zeners, and I also notice one of the 'fat' 0.2 Ohm resistors is high (0.4)
Yes, I am trying to talk myself out of replacing the bulk capacitors !!!
Tim


On Fri, 4 Jun 2021 at 23:24, walter shawlee <walter2@sphere.bc.ca> wrote:

Dave,

for clarity, when the Tek parts list says it is selected, it means that
Tek found the generic specs too loose to insure correct operation, so they
specified one or more parameters that had to be tested to insure that. In
the case of the MJE800, it is almost certainly beta, as the spec is
incredibly wide, and Tek will have wanted at least a threshold beta at a
specific current. All darlingtons have some issues with this parameter, and
this is why they are almost never used in audio circuits, the variation is
just too wide.

the Tek 151-spec for the transistors says exactly what the selection
criteria are, you can download the transistors specs from the Tekwiki to
get an exact answer. What is more puzzling is that transistor is hot,
unless there is an unexpected load on the downstream side, or collector
short to chassis ground, this probably shouldn't happen. Not so easy to
check a darlington with and ohmmeter, as it may have a CE shunt diode and
bias resistors, which really complicate things.

Are you sure there is no issue AFTER the transistor? Worth it to really
check, as bypass caps seems to be the eternal nemesis of all Tek hardware.
If the part really is bad, let me know, I will drop one in an envelope to
you for free.

all the best,
walter (walter2 -at- sphere.bc.ca)
sphere research corp.







Dave Peterson
 

My 5440 PS parts list. Let's see if copy-n-paste works from Excel:

Capacitor Value Tolerance Voltage Replacement Part Capacitance Voltage Tol Irip Life Lead Sep Dimensions
C800 170uF +50-10% 250v UCY2E181MHD 180uF 250v 20% 1110mA 7.5mm 31.5mm x 16mm
C848 10,000uF +100-10% 12v LGY1E123MELA40 12000uF 25v 20% 2600mA 10mm 40mm x 25mm
C875 7500uF +100-10% 25v LGY1H822MELB40 8200uF 50v 20% 2800mA 10mm 40mm x 30mm
C876 7500uF +100-10% 25v LGY1H822MELB40 8200uF 50v 20% 2800mA 10mm 40mm x 30mm
C930 5000uF +45-10% 50v LGY1H562MELB30 5600uF 50v 20% 2300mA 10mm 30mm x 30mm
C932 3000uF +100-10% 50v LGY1H562MELB30 5600uF 50v 20% 2300mA 10mm 30mm x 30mm


Diode Current Tek Part # Voltage Replacement Part
CR800 400mA 152-0107-00 200v NA - working
CR801 400mA 152-0107-00 200v NA - working
CR802 400mA 152-0107-00 200v NA - working
CR803 400mA 152-0107-00 200v NA - working

Rectifier Current Tek Part # Voltage Replacement Part
CR848 2.5A 152-0556-01 50v NA - working
CR875 2.5A SDA10271K 50v NA - working
CR930 1500mA 152-0488-00 200v KBP210G


Fuse Value Type Voltage Replacement Part Alternate
F300 1.25A 3AG Slow 250v 03131.25HXP 0034.5227
F410 0.3A 3AG Slow 250v 0313.300MXP
F800 0.25A 3AG Fast 250v 0313.250MXP


Other Qty Lead Separation eBay Item Number
0.75" Cap Adapter 2 0.75 in 272740198342
15.5mm Cap Adapter 1 15.5 mm 272777454683

I'm a bit of a Nichicon and Mouser fan. I enjoy their documentation. Helps with the learning.

Hope this helps anyone interested.
Dave


Michael W. Lynch
 

On Sat, Jun 5, 2021 at 12:29 PM, Dave Peterson wrote:


Other Qty Lead Separation eBay Item Number
0.75" Cap Adapter 2 0.75 in 272740198342
15.5mm Cap Adapter 1 15.5 mm 272777454683
Dave,

Note that Alex has changed the design of his capacitor adapter boards and they now also accept "Snap in" caps directly, provided they feature a 10mm lead spacing. This opens up more options for caps. I use a lot of NICHICON caps and MOUSER is my "go to" as they are just one or two shipping days away from my location.

--
Michael Lynch
Dardanelle, AR


Dave Peterson
 

To wrap this case:

I ordered a set of replacement MJE800 equivalents from Mouser for 6-bucks-n-change. Got a few nice Pomona jumpers with micro-grabbers to justify the shipping.

Walter suggested offline that my problem was more likely a short downstream. Guess what: Walter's a good guesser. The issue turned out to be a shorted 1uF tantalum decoupling cap on the main interface board on the -30v supply. As a reminder: the blazing hot MJE800 was the -15v supply. Odd, but lesson learned: interacting regulators may do unexpected things.

This was part of a electrolytic recapping and 152-0488-00 replacement. Not necessarily bad, but one of those, "while I'm in there I might as well ..." exercises.

With everything reassembled the PS came up fine on the bench with a variac and meters wired up to monitor ac and dc voltages and currents. Everything came up with no problem and supply current stayed down around 150mV. Regulated voltages were a bit over-range, but unloaded.

After reassembling the scope I was able to run through the calibration procedure. +/- 30v supplies came down to in-range and adjusted to within 10mv. Other supplies are now in spec. I now have a nicely working 5440 scope.

However, and I haven't figured out why yet, the original purpose of this scope was to run a 5CT1N curve tracer. The curve tracer works nicely (very sharp traces) in a 5111A, but in the 5440, which otherwise works at full BW with clean traces, produces fat ugly blurry curve tracer lines. WTF? A puzzle to be solved. Rather frustrating after all the work.

One final lesson learned in this exercise: plug-in curve tracers are relatively low power small signal devices. The Q880 that I thought was shorted was just under driven by the 5CT1N. It can only go up to 20mA Ic, but the MJE800 Darlington transistors drive 4A with several hundred mA Ibe. The plug-in can only put out 1mA steps at most. And being Darlingtons they also have an elevated Vbe, which seemed to not function at all at low Ibe steps. What I interpreted as a shorted collector-emitter was just the first step driving full vertical range. More careful examination of the MJE800 datasheet and curve tracer settings showed the transistor actually never even got out of saturation. To a 5CT1N it does kind of look like a short. Beginners, eh?

Thanks to all who helped, and hope my lessons learned help others down the road.
Dave


Harvey White
 

You might want to check on this, but the 5000 series plugins and the 5400 series plugins were not supposed to be interchangeable. I think you certainly couldn't use a 5400 series in a 5000, and I *think* that the reverse is true.

They are not the same scope series, from what I know.

Fat blurry traces indicates that the supply voltages are off, the HV is likely strange, etc.  Not sure, but I'd do a very close read on the manual.

Harvey

On 6/15/2021 1:25 AM, Dave Peterson via groups.io wrote:
To wrap this case:

I ordered a set of replacement MJE800 equivalents from Mouser for 6-bucks-n-change. Got a few nice Pomona jumpers with micro-grabbers to justify the shipping.

Walter suggested offline that my problem was more likely a short downstream. Guess what: Walter's a good guesser. The issue turned out to be a shorted 1uF tantalum decoupling cap on the main interface board on the -30v supply. As a reminder: the blazing hot MJE800 was the -15v supply. Odd, but lesson learned: interacting regulators may do unexpected things.

This was part of a electrolytic recapping and 152-0488-00 replacement. Not necessarily bad, but one of those, "while I'm in there I might as well ..." exercises.

With everything reassembled the PS came up fine on the bench with a variac and meters wired up to monitor ac and dc voltages and currents. Everything came up with no problem and supply current stayed down around 150mV. Regulated voltages were a bit over-range, but unloaded.

After reassembling the scope I was able to run through the calibration procedure. +/- 30v supplies came down to in-range and adjusted to within 10mv. Other supplies are now in spec. I now have a nicely working 5440 scope.

However, and I haven't figured out why yet, the original purpose of this scope was to run a 5CT1N curve tracer. The curve tracer works nicely (very sharp traces) in a 5111A, but in the 5440, which otherwise works at full BW with clean traces, produces fat ugly blurry curve tracer lines. WTF? A puzzle to be solved. Rather frustrating after all the work.

One final lesson learned in this exercise: plug-in curve tracers are relatively low power small signal devices. The Q880 that I thought was shorted was just under driven by the 5CT1N. It can only go up to 20mA Ic, but the MJE800 Darlington transistors drive 4A with several hundred mA Ibe. The plug-in can only put out 1mA steps at most. And being Darlingtons they also have an elevated Vbe, which seemed to not function at all at low Ibe steps. What I interpreted as a shorted collector-emitter was just the first step driving full vertical range. More careful examination of the MJE800 datasheet and curve tracer settings showed the transistor actually never even got out of saturation. To a 5CT1N it does kind of look like a short. Beginners, eh?

Thanks to all who helped, and hope my lessons learned help others down the road.
Dave





Dave Peterson
 

Thanks Harvey,

I did discover the plug-in incompatibility when I tried installing the 5440 plug-ins in the 5111A. Jeff Dutky pointed me to the pertinent section of the Tek Wiki that describes this.

In that I also found that the 5111A plug-ins are "forward compatible" and work in the 5440. I was a tad surprised to see the 5A26 from the 5111A gave readouts on the 5440. While tuning up the 5440, on a lark, I used the 5A26 from the 5111A and the 5A48 from the 5440 to prove I could get four channels showing at once with 5 readouts. The time base is a delayed 5B42, but can't do "dual" A/B sweeps. (That'd be 8 sweeps! :P) But it did prove the scope could do the 4 channels. FWIW. I don't have much experience with these scopes, so I don't even know if a dual delayed sweep time base is available or possible. I rather doubt it. So I don't think there's a lower-right quadrant readout to be had. I digress. Readouts are a whole other thread!

The calibration procedure in the 5440 manual is rather brief, but it does include a HV voltage check, which was within spec. The 5A48 need some very slight tweaking to bring CH1 & CH2 into agreement and get the step attenuator adjustments spot on. Once it was tuned it did very well with edge compensation and provides very nice sharp traces across the volts/div range. Noise, of course, shows up below 5mV/div. But that's to be expected (see my other posts about my large FM interference from local broadcast tower).

The horizontal cal seems a bit deficient, and the time base should be calibrated too. It also has some knob damage to the delayed time/div portion that I'll need to fix. And some switch maintenance. It'll need to be partially disassembled to get at them. These issues are minor and don't appear to affect the horizontal performance. The timing accuracy and stability are good.

Perhaps I need to check the 5CT1N supplies in the center slot of the 5440? Could it be a supply issue in the curve tracer? The cal does have one swapping the vertical plug-in back and forth between the left and right slots to verify vertical position and gain. And I ran through the volts/div ranges with a PG506 in both slots. Both producing accurate and sharp traces. But that doesn't guarantee the 5CT1N is happy, does it. The shorted tantalum decoupling cap was right next to the center slot, for what that's worth. I replaced it with a 50v tantalum with the same 1uF value.

I'll have to run through a 5CT1N check-out in the 5440, now that you've brought that up Harvey. Thanks for the suggestion.
Dave

On Tuesday, June 15, 2021, 07:22:50 AM PDT, Harvey White <madyn@dragonworks.info> wrote:

You might want to check on this, but the 5000 series plugins and the
5400 series plugins were not supposed to be interchangeable. I think you
certainly couldn't use a 5400 series in a 5000, and I *think* that the
reverse is true.

They are not the same scope series, from what I know.

Fat blurry traces indicates that the supply voltages are off, the HV is
likely strange, etc.  Not sure, but I'd do a very close read on the manual.

Harvey


On 6/15/2021 1:25 AM, Dave Peterson via groups.io wrote:
To wrap this case:

I ordered a set of replacement MJE800 equivalents from Mouser for 6-bucks-n-change. Got a few nice Pomona jumpers with micro-grabbers to justify the shipping.

Walter suggested offline that my problem was more likely a short downstream. Guess what: Walter's a good guesser. The issue turned out to be a shorted 1uF tantalum decoupling cap on the main interface board on the -30v supply. As a reminder: the blazing hot MJE800 was the -15v supply. Odd, but lesson learned: interacting regulators may do unexpected things.

This was part of a electrolytic recapping and 152-0488-00 replacement. Not necessarily bad, but one of those, "while I'm in there I might as well ..." exercises.

With everything reassembled the PS came up fine on the bench with a variac and meters wired up to monitor ac and dc voltages and currents. Everything came up with no problem and supply current stayed down around 150mV. Regulated voltages were a bit over-range, but unloaded.

After reassembling the scope I was able to run through the calibration procedure. +/- 30v supplies came down to in-range and adjusted to within 10mv. Other supplies are now in spec. I now have a nicely working 5440 scope.

However, and I haven't figured out why yet, the original purpose of this scope was to run a 5CT1N curve tracer. The curve tracer works nicely (very sharp traces) in a 5111A, but in the 5440, which otherwise works at full BW with clean traces, produces fat ugly blurry curve tracer lines. WTF? A puzzle to be solved. Rather frustrating after all the work.

One final lesson learned in this exercise: plug-in curve tracers are relatively low power small signal devices. The Q880 that I thought was shorted was just under driven by the 5CT1N. It can only go up to 20mA Ic, but the MJE800 Darlington transistors drive 4A with several hundred mA Ibe. The plug-in can only put out 1mA steps at most. And being Darlingtons they also have an elevated Vbe, which seemed to not function at all at low Ibe steps. What I interpreted as a shorted collector-emitter was just the first step driving full vertical range. More careful examination of the MJE800 datasheet and curve tracer settings showed the transistor actually never even got out of saturation. To a 5CT1N it does kind of look like a short. Beginners, eh?

Thanks to all who helped, and hope my lessons learned help others down the road.
Dave






Szabolcs Szigeti
 

Hi,

Be careful, the calibration instructions are dispersed in multiple places
in the manual, probably because the 5440 mainframe was originally a two
part unit, with the display and the mainframe section. To fully
calibrate, you'll need a special plugin, the 067-0680-00 calibration
fixture. So check all parts of the manual, not only the Adjustments section!

Indeed the plugins for the 5000 series can be used in the 5400 series, but
not the other way around. There are some 5000 series plugins with Readout
(no N in the model number), they will work in both, but obviously readout
will only be available in the 5440.

For the 5CT1N check there are two capacitors that in some units are under
rated and will cause a lot of interesting results and sometimes smoke. I
had to replace them in my unit, they were already blown when I got the
plugin. See: http://hakanh.com/dl/docs/hardtofind/CT1N.pdf

I actually very much like the 5440 scope, obviously no match for a 7000
series, but has a very nice large and sharp screen, and some fairly nice
plugins.

Szabolcs


Dave Peterson via groups.io <davidpinsf=yahoo.com@groups.io> ezt írta
(időpont: 2021. jún. 15., K, 16:52):

Thanks Harvey,

I did discover the plug-in incompatibility when I tried installing the
5440 plug-ins in the 5111A. Jeff Dutky pointed me to the pertinent section
of the Tek Wiki that describes this.

In that I also found that the 5111A plug-ins are "forward compatible" and
work in the 5440. I was a tad surprised to see the 5A26 from the 5111A gave
readouts on the 5440. While tuning up the 5440, on a lark, I used the 5A26
from the 5111A and the 5A48 from the 5440 to prove I could get four
channels showing at once with 5 readouts. The time base is a delayed 5B42,
but can't do "dual" A/B sweeps. (That'd be 8 sweeps! :P) But it did prove
the scope could do the 4 channels. FWIW. I don't have much experience with
these scopes, so I don't even know if a dual delayed sweep time base is
available or possible. I rather doubt it. So I don't think there's a
lower-right quadrant readout to be had. I digress. Readouts are a whole
other thread!

The calibration procedure in the 5440 manual is rather brief, but it does
include a HV voltage check, which was within spec. The 5A48 need some very
slight tweaking to bring CH1 & CH2 into agreement and get the step
attenuator adjustments spot on. Once it was tuned it did very well with
edge compensation and provides very nice sharp traces across the volts/div
range. Noise, of course, shows up below 5mV/div. But that's to be expected
(see my other posts about my large FM interference from local broadcast
tower).

The horizontal cal seems a bit deficient, and the time base should be
calibrated too. It also has some knob damage to the delayed time/div
portion that I'll need to fix. And some switch maintenance. It'll need to
be partially disassembled to get at them. These issues are minor and don't
appear to affect the horizontal performance. The timing accuracy and
stability are good.

Perhaps I need to check the 5CT1N supplies in the center slot of the 5440?
Could it be a supply issue in the curve tracer? The cal does have one
swapping the vertical plug-in back and forth between the left and right
slots to verify vertical position and gain. And I ran through the volts/div
ranges with a PG506 in both slots. Both producing accurate and sharp
traces. But that doesn't guarantee the 5CT1N is happy, does it. The shorted
tantalum decoupling cap was right next to the center slot, for what that's
worth. I replaced it with a 50v tantalum with the same 1uF value.

I'll have to run through a 5CT1N check-out in the 5440, now that you've
brought that up Harvey. Thanks for the suggestion.
Dave


On Tuesday, June 15, 2021, 07:22:50 AM PDT, Harvey White <
madyn@dragonworks.info> wrote:

You might want to check on this, but the 5000 series plugins and the
5400 series plugins were not supposed to be interchangeable. I think you
certainly couldn't use a 5400 series in a 5000, and I *think* that the
reverse is true.

They are not the same scope series, from what I know.

Fat blurry traces indicates that the supply voltages are off, the HV is
likely strange, etc. Not sure, but I'd do a very close read on the manual.

Harvey


On 6/15/2021 1:25 AM, Dave Peterson via groups.io wrote:
To wrap this case:

I ordered a set of replacement MJE800 equivalents from Mouser for
6-bucks-n-change. Got a few nice Pomona jumpers with micro-grabbers to
justify the shipping.

Walter suggested offline that my problem was more likely a short
downstream. Guess what: Walter's a good guesser. The issue turned out to be
a shorted 1uF tantalum decoupling cap on the main interface board on the
-30v supply. As a reminder: the blazing hot MJE800 was the -15v supply.
Odd, but lesson learned: interacting regulators may do unexpected things.

This was part of a electrolytic recapping and 152-0488-00 replacement.
Not necessarily bad, but one of those, "while I'm in there I might as well
..." exercises.

With everything reassembled the PS came up fine on the bench with a
variac and meters wired up to monitor ac and dc voltages and currents.
Everything came up with no problem and supply current stayed down around
150mV. Regulated voltages were a bit over-range, but unloaded.

After reassembling the scope I was able to run through the calibration
procedure. +/- 30v supplies came down to in-range and adjusted to within
10mv. Other supplies are now in spec. I now have a nicely working 5440
scope.

However, and I haven't figured out why yet, the original purpose of this
scope was to run a 5CT1N curve tracer. The curve tracer works nicely (very
sharp traces) in a 5111A, but in the 5440, which otherwise works at full BW
with clean traces, produces fat ugly blurry curve tracer lines. WTF? A
puzzle to be solved. Rather frustrating after all the work.

One final lesson learned in this exercise: plug-in curve tracers are
relatively low power small signal devices. The Q880 that I thought was
shorted was just under driven by the 5CT1N. It can only go up to 20mA Ic,
but the MJE800 Darlington transistors drive 4A with several hundred mA Ibe.
The plug-in can only put out 1mA steps at most. And being Darlingtons they
also have an elevated Vbe, which seemed to not function at all at low Ibe
steps. What I interpreted as a shorted collector-emitter was just the first
step driving full vertical range. More careful examination of the MJE800
datasheet and curve tracer settings showed the transistor actually never
even got out of saturation. To a 5CT1N it does kind of look like a short.
Beginners, eh?

Thanks to all who helped, and hope my lessons learned help others down
the road.
Dave















Mark Vincent
 

Szabolcs,

I already let Dave know about the two 100mfd 25V in the collector circuit of the two push-pull transistors on the primary of the transformer. I cannot understand why a 25V was used when the B+ is 30V. The 240 ohm resistors from B+ to the condenser and collectors will not drop enough voltage to keep the condenser under 25V. I suggested using a ULD 50V as Mouser has these in stock at this time. The 35V types will be in later this year. These must be the two you refer to in your post.

Mark