Repairing a Siemens B1045 Multimeter

[david77]

Yes, I did it again. Just bought another bench top multimeter off ebay. Of course one that was sold as broken, in the past that has always worked out. I always managed to get them back to life, just trivial failures.

But first things first. The B1045 is quite a nice meter, 5 1/2 digits, auto ranging V-Ohm-Amps, 4-wire ohms function. The smallest range is a bit unusual, it is 0.2, so that effectively gives you a resolution down to 0.00001V, very nice indeed.
This meter also appears to be a genuine Siemens design, not like their handhelds from the same time that were made by Norma in Austria.
My unit was probably made sometime between 1979 and 1981. They are rather rare, they don't show up much either on ebay or at the various used test gear sellers. And if they do and are in working condition they fetch relatively high prices.

I got my meter yesterday, tried it out immediately and found it really didn't work. It only shows random numbers on the display and has some sticking buttons, cosmetically at least it's good.
Tonight I opened it up to have a quick look inside. Nice design with real computer contol, there's a 8085 SBC inside doing all the heavy work. I'll talk about that in a future post, also photos will come later.

The reason I'm writing this now is because I'm somewhat pi**sed off. Somebody's tried to repair it before me and they really made a mess of things. This blithering idiot just bridged a 7812 regulator, thus feeding unregulated 22V into a big part of the circuit . Needless to say there are plenty of 4000 series in it. As we all know they're only happy up to 18V. Also needless to say they are not socketed. I pulled two at random and tested them, both fried, of course.

This numpty tried to find a fault in the PSU region, cutting traces willy-nilly, most of them in two places . Instead of bridging the cuts with a short piece of wire, they created a rats nest of hook-up wires on the bottom of the PCB . I had to sort all that out before I could even start my own fault finding. While knitting this mess of wires the idiot jumped across the regulator and made things 10x worse.
At least the SBC is still ok, thanks to its separate 5V supply.

[Vgkid]

Good luck with the repair, do post pics.

[retiredcaps]

they created a rats nest of hook-up wires on the bottom of the PCB

Before and after pictures would be appreciated.

[Napalm2002]

Hell yeah very intrresting'!

[Fraser]

If there is one thing I really hate, its when an unqualified person has already had a go at a piece of faulty equipment. 'Virgin' equipment is usually pretty 'honest' in that it once worked and then something went wrong . Equipment that has already been 'got at' can end up with a multitude of additional faults introduced by the previous 'tech'. I have discoved chips inserted in sockets, or soldered into PCB's reversed  and components that had been replaced by 'compatable' parts that weren't ! Cut traces are an awful discovery as well..... especially if not done in a professional manner. The single worst discovery I made was an expensive bit of kit that someone had just gone around replacing chips in. Fine to a point, but not when they replaced a MPU that contained ROM. The new part was without programming  . That unit ended up as scrap.

I love to see faulty kit with its original calibration seals intact.... its a Virgin 

[david77]

Some more news on the meter.
First of all let me show you some pictures.

The outside: Aluminium enclosure, trendy brown colour and very deep. Relatively clean and in general in good condition.



Now with the hood up.



On the lower right side you can just about see the +/- 12V regulators, both of wich were dead.
To the left of them there's the whole bank of relays for the input selection/attenuation. The right-most relay on the piece of perfboard is a bodge that I made. There used to sit a reed relay like the others around it which was shot. It's a DPST  type and obviously I didn't have one on hand so made that adapter. More on that later. Just off the centre to the left are two relays that also sit on a little adpter board. I have no idea why or if it is original. There are three identical footprints on the main board, one is populated by the proper relay for the footprint the other two sit on that board. Repair maybe? Beneath this adapter board lies the attenuator resistor network, a FN207 made by a company called A-B.

The cards from top to bottom.
- The 8085 controller board, ribbon cables go off to the frontpanel.
- What I'd call the main A/D converter board, A/D conversion is done by LD120/LD121.
- The TRMS converter board, apparently an option not fitted to all of these meters.
- Not quite sure about that one, probably the resistance range, might do more as it seem a bit too complex for only measuring resistance.



At the moment I'm a bit confused by the unit. There is evidence of heavy modifications and I dont know who did it or why they did it. I'm pretty sure it's been done by one person or organisation, might even have been done by Siemens themselves.
There's the thing with relays I mentioned above but there's more. On the mainboard there are a few bodge wires and even what looks like a couple of bodged in parts. There are a few of those on the plug in boards as well. It looks like somebody's drilled the PCB's and added proper solder pads for those bodged in parts. I have no idea if that is possible, but it shure looks like it.



Here on the AD card you see the little blue trimpot in the top left corner. It uses one of the fiducials and two holes with proper pads around them. The three connections to the rest of that board are made using wires. The pot sits on top of a QC sticker.



Then there's the relay on the Ohms board, you can clearly see the footprint of the original part and what is in there. It's got proper solder pads, too. Not sure whats going on there  .
The wires used to wire in all the extra components and bodge connections are all off the same type, so chances are it was all done by the same person.

The only visible damage I could find so far was this blown track (centre of image): There was some scorching around the relay contact next to the solder lug. Look at the PCB next to the solder lugs, the trace ran from pin 8 of K8 to the via on the right of the two solder lugs. It looks like there was a flash over from the relay to this solder lug, the jumper wire between the lugs lets you disconnect the -12V rail.



The missing piece of trace was vapourized. These reed relays are used for the mA ranges, so I suspect someone exceeded the max. allowed voltage, that created the flash over and killed the PSU.



That's the back of the mainboard right under the supply section. You might be able to see some of the double cut tracks I had to repair. I've not taken a pic of the before state.
These same green wires are used on all the mods in this unit, they were glued to the PCB with hot glue once. Most of the hot snot has come off the board over the years.

I've now replaced the 4000 series IC's on the ADC board, they were all blown but that hasn't really changed anything. I still get random numbers on the display. There is some reaction to an applied voltage on the inputs but not really predictable.
If somebody has any idea how to check if the LD120/121 are fried please do leave it below, anyway I'm expecting a quote for a set of them.
The PSU looks fine now, I'm confident that problem is solved.
The manual is on its way, can't really continue doing anything sensible without that.

[david77]

It's been a while but I've not given up yet.

I did receive the manual, it contains the circuit diagrams but sadly no detailed explanation how the meter works and also no waveforms or what voltages to expect at the test points.
At least the modifications I found are also in the diagrams, so that's one mystery cleared up.

The datasheet of the LD120/121 refers to a an app note AN77-3 in which interfacing it to a µC is described in more detail, that would be very interesting. I've looked through all the interwebs but can't seem to find it. Maybe someone's got an old Siliconix data book with it in it?

What the meter does at the moment is it shows random numbers in all settings, the display changes wildly. I have pulled all the cards except the processor board and the A/D converter. In this setup at least the VDC range should still work.
I've found that in the 2 VDC range I can get the meter to display things in a predictable fashion. If I apply about +1,6V to the input the meter reads 0,0000 Volts - not very stable but recognisable.
At about +1,1V it shows overrange, at +2,6V it reads +1,00V and at +2,9V it reads +1,2632 and stays there even if I increase the input voltage further.
In all other VDC ranges it's all over the shop, no pattern I could find.

I still suspect the A/D chipset to be faulty. I've ordered the LD120/121 from the UK and have already received the LD120.
Sadly that IC was apparently DOA. I unpacked it from its static baggy - wearing a wrist strap! - and plugged it in only to find the meter now constantly reading zero. The LD120 got unnaturally hot, all it's outputs being pulled to -0,5V. I swapped the old one back in and we were back to random numbers. I'm still waiting for the LD121, the digital part of the A/D converter chip set.

I've also replaced most of the 4000 and 74 logic on the A/D board and the main board, as those could have been affected by the +12V rail having been at +20V.

At this stage I'm out of ideas. Any suggestions?

Attached are the diagrams of the
-A/D Converter
-Main board
-Processor board.

[david77]

Let me try to talk you trough it. At least as far as I think I understand what's happening.

The main board, Picture 31:

On the lower left is the input attenuator. I've checked that and it works correctly.
Lower right is the PSU. All voltages present and correct.
Top left is the relay driver circuitry. As far as I can tell that all does what it's supposed to.
Top middle and right are the opto isolators that connect the processor board to the rest of the system. I've replaced all the 74 logic there as it might have gotten some of the +20V before I repaired the PSU

The A/D Converter, Picture 37:

Lower left and middle. The lines designated PLL~ carry 50Hz AC voltage, which gets rectified and banged into shape by V2 and V1. This 50Hz square wave is fed into a 4046 PLL (V5) which converts it up to 204,80kHz. This is the clock signal for the LD120/121 pair.
Top left. V17 is a FET switch (LF13202) which receives the output of the input attenuator and routes it to the LF356. It also periodically switches to signal ground, apparently for the auto zero function. The LF356 can either operate as a buffer (v=1) or as an amplifier with a gain of 10. This is determined via the X1/X10  signal and V13,14. V14 is a 4066. After this stage the signal goes off into the LD120 which does it's thing. Here's where my understanding gets a bit sketchy.
Apparently Siemens decided not to use the LD120's internal buffer.
All of the aforementioned stages of the circuit do seem to work as they should. I checked it all and can't find a fault there  .

[david77]

After a few weeks of struggling to get the needed components the Siemens B1045 meter is finally back in working condition  .

The last missing part was the LD121ACJ, the digital controller for the LD120 A/D converter. It finally arrived from the UK in todays post, I put it in and voila the meter was back.