What do you want to see?

[tekfan]

I am constantly getting new (well old) gear and can't find any good high-res pictures so I'll upload some. and write some accompanying text. I hope someone will find it helpful.
Please tell what you want to see a teardown of first. I'll try to do them in the order that you want. I hope I can do a teardown each week. Since hacklordsniper started the same thing I tought why don't I do the same.

These are the teardowns I can provide:

Tek 7905 (500Mhz mainframe)
Tek 7603 (100Mhz mainframe)
Tek 7613 (100Mhz analog storage mainframe)
Tek 561 (tube 10Mhz, sadly missing HV transformer)
Tek 2236 (100MHz scope with counter timer multimeter)
Tek DSA602A (digital 1GHz, 2GS/s mainframe)
Tek 1502 TDR (sady doesn't work, tunnel diodes faulty)

Multimeters:
Datron 1072 (7.5 digit)
Solartron 7150 (6.5 digit)
Solartron/Schlumberger 7075 (7.5 digit)
Ballantine 321 (TRMS VTVM)

Signal sources:
Tek FG504 (40Mhz function generator)
Tek 106 pulse generator (<1ns risetime)
Tek 191 sig gen (50kHz-100MHz)

Tektronix Plugins:
7A24 (400MHz vertical amp)
7A26 (200MHz vertical amp)
7A18 (80MHz vertical amp)
7A13 (100MHz differential amp)
7B50 (100Mhz time base)
7B53 (100Mhz time base)
7B10 (1GHz time base)
7D01 (logic analyzer and display formatter)
7D15 (250MHz counter)
11A32 (400MHz vertical amp)
11A71 (1GHz vertical amp)

Power supplies:
Lambda LP-522-FM (40V 1.8A)

LCR meters:
ESI 2160 LCR videobridge

Other:
Tek TM501
Tek TM504
Tek AM501 (current probe amp)
Bruel & Kjaer  2305 strip chart recorder (all tube)

[saturation]

WOW, that's a whos- who of top gear.  Did you just get free those 7.5 digit DMM?

I'd love to see these; but more I'd like to know if they still work!

Datron 1072 (7.5 digit)
Solartron/Schlumberger 7075 (7.5 digit)

[tekfan]

OK. First up is the Schlumberger/Solartron 7075 Digital multimeter.
I got it less than a week ago and it is in need of repair and calibration (that's what you get for 30USD on ebay). The basic functions work but there is too much DC offset on the lowest range. I don't have a service or cal manual so I'm just fiddling around with the trimmers.
This particular one is from 1977. It has several integration times (from 1ms which gives you 1400 count resolution or you can select 10s time with 14000000 count resolution). On the 1ms time you get 1000 readings per second so I would imagine that would be very useful for fast datalogging.It has built in self test features, DC volts, true RMS ACV range, four wire resistance measurment, ratio measurments, GPIB interface (oh yeah not even RS-232). The input connector is a rather strange one. Input resistance on the 10 volt range is 1Tohm (that's terra ohm), 200Gohm on 1 volt range, 20Gohm on 100mV and 15Gohm on 10mV. And you tought the new handheld meters have a high input resistance. Although this thing is almost 50cm deep and weighs 10kg.
I can't say much else about this meter before It is repaired and cal'd.
More pictures can be seen at: http://s1138.photobucket.com/albums/n537/tekscopes/

On with the pictures:
This is how it looked like when I received it. It's had a hard life.



After a day of cleaning this is the result:



Inside


This is the ohms section. The precision four wire resistors can be seen and the big blue thing is a 8.19meg 0.05% resistor


This is the reference 7 volt zener diode. It's made by TI. The two wires on the bottom of it are the heater connections.
This heats the diode and brings it to it's operating temperature.


Input section, HF compensation capacitor, high voltage resistor, some input protection can be seen.


Optical chopper assembly. Four LEDs shine on four LDRs. This allows for a very low input offset voltage.


Optical chopper driver


This is the reference oscillator. The teflon adjustment trimmer can be seen that is used for setting the frequency to exactly 200kHz.
I wonder why they haven't used a standard quartz crystal. The frequency was still spot on when I got the thing.


True rms converter. Big black box of mysteries. The input is squared, averaged across the big 10uF capacitor and then the square root of that value is calculated. This could possibly be replaced by a AD536 or AD637 DIP14 chip.

[saturation]

Awesome, what a collection of gear you have on the table.  More later, lots to digest.

[Russel]

Wow! You did an incredible job of cleaning up that Schlumberger/Solartron 7075 Digital multimeter! If you feel so inclined, it would be very interesting to know how you cleaned it so well.

[tekfan]

I usually shower the boards with water and get the dirt off with a toothbrush. Window cleaner is also nice for washing external parts such as knobs. The IC sockets on this one were a bit tricky since they hold a lot of water. I don't have a compressor handy but a hair dryer works very well. I've also used an oven (not microwave of course) set to about 50°c and it works also very well at drying the boards.

[dimlow]

I find the dish washer works well, but don't have salt or detergent in it as it can take off more than just dirt. Also dont use the drying in the washer it make things rust. One more thing ,use the middle setting for heat.

I do this for some of the old Micro's I collect, they come out looking new.

[alm]

Any issues with meeting the input impedance specs after washing? Any leakage due to residue left on the PCB will swamp it.

[tekfan]

Yes. I would imagine the Input impedance drops quite a bit. It can be minimised by washing the input section with denatured alcohol. It's nice to have such sensitive equipment in a dry area but then you get all sorts of ESD problems. Anyway I don't need such a high impedance when doing measurements. I still think it's amasing that they could accomplish something like this 35 years ago. The digital circuits almost exclusively use 7400 series logic.

[Alex]

Hi, assuming that the stuff on the boards is just dust, how about blowing compressed air (for a reasonable distance not to cause damage)?

Apart from impedance changes, I am worried that liquids might sneak into non-sealed components, like that relay.

I used to clean circuit boards I collected from TVs (CRTs) people threw away on the street and clean them off with water and soap. I did not expect anything to work afterwards, I connected the chassis to a 4.5V battery (remember those flat ones with the huge spring tabs?) and the other pole to a small lightbulb and probed around the circuit board looking for changes in light output. Good times.

[tekfan]

I noticed that moisture actually got into one of the relays. You can see the condensation in one of the pictures. I just took the top off and aimed the hair dryer at it for a couple of mins. Once the water evaporates there's really nothing to worry about.

You might want to look at this site to see how this guy cleaned and repaired his dirty receiver:
http://www.amplifier.cd/Verstaerker/Reparatur/Reparatur_Receiver.html

Also check the site out:
http://www.amplifier.cd/

More reviews coming soon!

[tekfan]

Next review / teardown.

This is the Datron 1072 Autocal Digital Multimeter. Back in the day Datron was a part of Wavetek. The meter is a 6.5 count one (switchable to 7.5 count with the input filter key). This has been my main meter for the past year and it's really useful when calibrating stuff. This one dates back from 1987 and has the following options: four wire resistance measurment, 1 MHz true RMS converter, current measurment, rear inputs, rear ratio inputs, GPIB. When i got it, it was set for 120 VAC operation. It came from the US and was used by Rockwell (the plane company) and it somehow found it's way to Europe. To change the mains voltage one must change a couple of jumpers inside. The whole thing is microprocessor controlled so you can do some pretty fancy stuff with it. You can enter some value and get a % deviation indication (very useful for resistor fabs), divide by x amount, subtract by x amount, measure a resistor and directly display the current flowing trough it. Then there is standard stuff such as min, max, average. The ohms guard terminal can be used to get a very accurate measurment of a resistor in circuit by nulling out the values of other reisistors. On the lowest DC volts range you get a resolution of 10 nano volts! Be on the lookout for these meters on ebay (models 1082, 1081, 1072, 1071, 1065, 1061). They go for insanely low prices so I suggest you pick one up if you have the room for it. See what options the meter has because if it doesn't have any you can just measure DC volts.

On with the pictures:

It's showing a voltage when nothing is connected because of the high impedance input's being charged up by static.


Front panel removed showing the plasma discharge display, keys and input terminals


Analog main board (top side of meter)


Two temperature compensated zener diodes for the reference voltage (not very impressive but it works)


Precision voltage divider resistors for the 100 and 1000 Volt ranges


AC board, used for AC voltage or AC current conversion to DC voltage


High frequency compensation network (reed relays, teflon trimmer capacitors)


Custom true RMS converter IC


Resistance measurment current generator board


Current board, precision four wire resistor can be seen


Driver board for the plasma discharge display, high voltage driver ICs (170 volts just for the display)


Optocouplers for isolation between the digital and floating analog boards


GPIB / IEEE-488 interface board


Digital board, memory keep alive battery can also be seen (still at 3 volts - they sure don't make them like they used to), the transformer can be seen in the upper left corner - these are actually two completely separate torodial transformers bound together, one for the digital and one for the floating analog section)


Main microprocessor


Quartz oscillator for the clock


Back of the thing


Suggestions for future teardowns would be much appreciated.

[alm]

Nice pics!

When i got it, it was set for 120 VAC operation. It came from the US and was used by Rockwell (the plane company) and it somehow found it's way to Europe. To change the mains voltage one must change a couple of jumpers inside.

Note that it requires a different crystal frequency (and some caps) depending on the mains frequency (50/60 Hz). It may need to be calibrated afterwards (I believe the official recommendation is to do a performance verification and do calibration if it fails). The only consequence of wrong frequency is reduced CMRR at your mains frequency.

See what options the meter has because if it doesn't have any you can just measure DC volts.

Options also depend on the model. I think the 10x2 is the same as the 10x1 model with some options as standard. The 108x series has better accuracy than the other series, but less features (eg. no current, limited math).

Two temperature compensated zener diodes for the reference voltage (not very impressive but it works)

I was also surprised by this, since most of their competitors use fancy heated zeners. They do seem to get better stability than all(?) the LM199/299/399 based equipment out there (at least in the 108x series), though, so they must be doing something right.

Digital board, memory keep alive battery can also be seen (still at 3 volts - they sure don't make them like they used to)

Note that these batteries will stay at almost the same voltage until they die, so it's hard to judge the condition. Since empty battery means loss of calibration constants, I would be conservative with estimating battery life, ie. follow the manufacturers recommendation of replacement every 10 years.

Back of the thing

Are those rear connectors standard and easy/affordable to acquire?

[tekfan]

Note that it requires a different crystal frequency (and some caps) depending on the mains frequency (50/60 Hz). It may need to be calibrated afterwards (I believe the official recommendation is to do a performance verification and do calibration if it fails). The only consequence of wrong frequency is reduced CMRR at your mains frequency.

I know! In the user manual it just says ''for line voltage and frequency conversion refer to service manual''. Sadly I don't have a complete service manual. It seems to still be very accurate. CMRR might be a problem on the lowest ranges but I usually don't use them. The thing also monitors the frequency when doing AC measurments. It might be doing the same with the line frequency although I doubt it.

Note that these batteries will stay at almost the same voltage until they die, so it's hard to judge the condition. Since empty battery means loss of calibration constants, I would be conservative with estimating battery life, ie. follow the manufacturers recommendation of replacement every 10 years.

It would suck to loose cal constants. Can you replace the battery by connecting a power supply to the meter when removing the battery? Anyone know where to get such batteries?

Are those rear connectors standard and easy/affordable to acquire?

I seriously don't even know what they're called. They look similar to Amphenol's PT series but are threaded and not bayonet locking. I suppose they chose these connectors for a reason. I was thinking of replacing them with DB-9 connectors.

Any suggestions for future teardowns?

[Alex]

My vote on the Bruel & Kjaer  2305 for something different.

[alm]

It would suck to loose cal constants. Can you replace the battery by connecting a power supply to the meter when removing the battery? Anyone know where to get such batteries?

I think there's a diode in series with the battery, which is reverse biased by the 5V supply while the unit is powered on. I would make sure that my soldering iron is not grounded (eg. by unplugging it before touching the joint and relying on heat capacity), because I believe one terminal of the battery is grounded. This is a 3.6V 1/2 AA lithium battery if I remember correctly, should be available from any decent electronic component distributor. Make sure to get one with leads attached.

I seriously don't even know what they're called. They look similar to Amphenol's PT series but are threaded and not bayonet locking. I suppose they chose these connectors for a reason. I was thinking of replacing them with DB-9 connectors.

They are by Pye Connectors, part no M7P/M7S (plug and socket) according to the manual, not sure how helpful this since Pye doesn't appear to exist anymore. I don't think DE-9 would be a good choice. Parameters I would expect to be important is contact resistance (especially in 2-wire resistance or current mode), high impedance (input impedance is >10Gohm up to 20V) and rated for voltages up to 1000V and whatever the current limit is. They may also have been designed to minimize thermal EMF. DE-9, which is just a low-speed digital connector which appears to satisfy none of them.

[Ferroto]

You peaked my interest with the socketed MC6800P. http://en.wikipedia.org/wiki/MC68000

I would so temporary remove that and do a 6800 computer

[tekfan]

Next is the Bruel & Kjaer 2305 Level recorder

Bruel & Kjaer were and still are probably the most renowned company for low frequency and audio measuring equipment (precision microphones, vibration transducers, vibration amps audio voltmeters etc.). They are still located in Denmark.
The 2305 is basically just a chart recorder. I got it at a radio/TV station auction (with a truckload of other stuff). It dates back from the late 50's and early 60's. It's all tube based except for the power supply rectification which uses diodes. The whole thing is very very very heavy and not exactly practical if you don't have a lot (and I do mean a lot of space). But it still shows what could be accomplished with nothing but tubes as the active components.

Shot of the whole thing.
A nice option is the amplifier response switch which allows true RMS, peak, average and DC coupled operation.
The writing speed switch determines the max speed that the pen can move at (selectable from 4 to 2000 millimeters per second)
The paper speed can also be set by means of a knob that connects to a gearbox inside. Adjustable from 300 nanometers to 100 milimeters per second. (300 nm/s is about 2.6 cm/day)


Closeup of the pen mechanism. The pen travel can be halved by simply moving the shaft to one of the two holes above the pen.


In the back the 10 tubes and the hughe power transformer (with many many taps and different windings).
I've also added a modern IEC power connector since it had the old style American 3 prong one.


The potentiometer is used as position sensing of the pen. This one is logarithmic. It can be replaced with a linear one that then gives the pen movement a linear response.


On the other side of the potentiometer the contacts can be seen. At first glance it looks like a wirewound resistor


Once opened up you can really see what's inside. Whoever designed this was probably a bit crazy.


Each junctions at two resistors are connected to a row of very fine sheets of metal The wiper then moves across these.


The giant linear actuator. Inside there is a former with two coils. One is the power coil used for moving the shaft back anf forth and the other is the feedback coil used for frequency and transient compensation. They are both surrounded by very powerful magnets. Similar to a speaker This technique is used to achieve true RMS measurments. Kind of like a standard moving coil meter. The actuator weighs about 7 kilos and makes up about a quarter of the weight of the entire recorder.


Actuator mounted back in after cleaning. The square silver can is the mechanical chopper running at 100Hz. This is part of the low offset amplifier.


Another shot of the chopper. The motor in the background is for the paper feed.


Gearbox for reducing the motor's revolutions. The leftmost gear spins the fastest and the rightmost the slowest. There is a gear that can be moved along these gears and that drives the teeth that move the paper forward.


Power supply filter capacitors and filter choke. After 50 years they still read reasonably good on the LCR meter (ESR is also good). I've put some new capcitors in parallel with these to reduce the ripple.


More transformers. The one in the front is probably the input transformer. I've seen a lot of audiophiles buy JS transformers for quite a bit of money. I'm not an audiophile and I'm proud of it.


I'll have a lot of time next week so suggestions for future reviews are welcome.

[saturation]

These are beautiful instruments!  Thanks for sharing the imagry.  I wish I had time to comment more on the internals you show.  Please keep them coming.

[benjius]

Hi Ferroto,

You have a great collection ! and you're a very good photographer, the macros are awesome !
Thanks a lot for shearing !

Ben.

[tekfan]

This is a review / teardown of the Tektronix FG 504 function generator. This one dates from 1980. It's meant to be used with the Tek TM500 series mainframes.
This function generator is the highest bandwidth fully analog function generator I have ever seen. It goes from 1mHz to 40 MHz. Has the usual stuff such as sine, triangle, square, symmetry/duty cycle (10-90%), offset. The things that not many function generators have is AM and FM modulation inputs, variable rise time (minimum is 6ns), sweep function (logarithmic and linear sweep) making it very useful for plotting amplifier responses on to a chart recorder. It also has various trigger functions like TRIG-will generate one cycle of a selected waveform, GATE-will continue making cycles when the trigger input is high, LOCK-can be PLL locked with an external source. One of the frequency ranges can be user defined by changing a capacitor inside (standard configuration is 20-20000 Hz for audio). The sine wave function has a distortion of less than 0.5% up to 100KHz so it's quite useful for amplifier testing.

Any future teardown requests?

Pictures:

Front panel showing all of the controls


The whole thing is very compact for such a high spec function generator. Everything is tightly packed onto 3 boards.


Main board - output amplifier, attenuator, rise/fall time switch.
Everything is gold plated hence the yellowish pictures (not a white balance problem)


Detail of output amplifier - the two big silver transistors have been replaced since the leads were broken on the original ones.
A lot of trimmer capacitors for high frequency compensation in this area.


Detail of the 50 ohm attenuator. The ceramic dividers are switched in or bypassed by means of the gold plated cam switches.
The 0.3 amp output fuse can also be seen. The generator provides almost 0.3 amps in short circuit.


One of the ceramic dividers lifted to show what's underneath. This is a 10dB divider.
The marks from the laser trimming process can be seen.


This is the function generator board that synthesizes the triangle and sine out of a square wave. Lots of diodes and custom ASIC's.


Timer board - lot's of frequency and symmetry adjustments


Detail of high frequency construction at the time


Nice clean step response from the function generator (terminated to 50 ohms) set to square wave.
The rise time is slightly less than 6ns. Measured with a 400MHz vertical amp so it's risetime is negligible.

[krivx]

I realize this thread has been quiet for some time but I thought I would revive it on the chance we could see more restorations. Please ?

[chrome]

Tek 7613 (100Mhz analog storage mainframe)

Solartron 7150 (6.5 digit)

Tek 106 pulse generator (<1ns risetime)

ESI 2160 LCR videobridge

Tek TM501
Tek TM504
Tek AM501 (current probe amp)

[free_electron]

wow. an lm301 opamnp in metalcan to99 .. made by AMD ...

[Rerouter]

tek tm501, mine came with a few options so would be nice to see whats different