Revox B 795

[markthefixer]

Ok, status:

ALL chips now socketed. Power supply moved forward to ease motor driver board external access with motor cable connected. Arm board power disconnected to allow motor to free run. Several 28 gauge insulated and stranded jumper wires were connected to IC 10 pins 1 & 7 for easier and more reliable probing on the board. Motor board cable limiting factor now.

Manual (or the printed sheet) is INCORRECT in original polarity directions for C15 and C23. Ohmmeter verification of connections confirmed pcb vs manual discrepancy. Corrected.

All original chips reinstalled - Robie's observed symptoms of motor "snapping" (slowly but there) to particular positions observed. Was "progress". The effect was due to one channel of IC 10(.2) pin 7 being zorched and output at negative (iiac) rail. Chip replaced temporarily with LM358 dual op amp.

ALL FOUR Hall outputs observed at the connector to be around 6 volts dc. Scoping pins while rotating showed that three hall outputs varied under 5 mv AC as observed with 1: 1 scope probe on Tektronix 465 scope on max AC sensitivity. One did not appear to move at all. But at that level of sensitivity and noise, minimal movement or none is hard to discern. I'm going DC with my HP 3468B 5.5 digit DMM.

Reading voltages differentially with a 3.5 digit DMM while slowly rotating turntable by hand showed a differential voltage on the J3 5(blue) to J3 4(violet) of minus 8 millivolts to plus 7 millivolts, with 6 full cycles per rotation (-7 to 0 to +8 to 0).
The output of IC 10.1 pin 1 is a differential amp with gain of 30.2, it's output ranged from +0.168v to -0.290v, so dividing by the gain gives -9.6mv to +5.5mv - so the op amp had a little offset voltage .. 2mv or so.

The OTHER side showed a differential voltage on the J3 3(gray) to J3 2(white) of plus 4 millivolts to plus 7 millivolts, again with 6 full cycles per rotation .
The output of IC 10.2 pin 7 is also a differential amp with gain of 30.2
it's output ranged from +0.055v to +0.165v, so dividing by the gain gives 1.8mv to 5mv.

So while one leg of the hall inputs DID show DC bias, it's signal component was either very low or missing.
The other three inputs are of such a low magnitude that the op-amp's output was insufficient to cross the diode gap and definitely not enough to bias the driver transistors on.

The diodes were tested and found to be functioning correctly (Vf ~0.6v), as were the driver transistors (B&K 510 test, out of circuit) . Motor windings were found to be about 28 ohms each. Assuming +/- 12v (unregulated) yields expected maximum motor current of half an amp (500mA).

some 5.5 digit voltage data for the two hall effect channels:
one sensor
J3 pin 5 blue wire
J3 pin 4 violet wire
other sensor
J3 pin 3 gray wire
J3 pin 2 white wire

J3 pin 8 orange wire = sensor power
J3 pin 7 yellow wire = motor ground & sensor ground.
J3 pin 1 black = motor winding
J3 pin 6 green = other motor winding

J3 pin 10 red = tach sensor
J3 pin 11 brown tach sensor

white V gray V gray-white Vdiff IC10 pin 7
6.0596 6.0554 0.004121 0.0541 (minimum)
6.0624 6.0547 0.007698 0.1610 (maximum)

violet V blue V violet-blue Vdiff IC10 pin 1
6.0886 6.0971 -0.008122 -0.284.3 (minimum)
6.0985 6.0985 +0.006778 +0.1663 (maximum)

At this point I believe the Hall effect outputs are insufficient to drive the circuit. Either the sensors are damaged, or the magnetic field has diminished. Perhaps the table was dropped and the gap between the ring magnet and the sensors opened up. The DC readings relative to the position of the turntable (I mapped them) were remarkably repeatable, showing the slight difference of flux intensity in the 6 magnets that make up the ring.

The small voltage difference between each of hall element's two outputs, especially relative to the hall power of 6.12v is troubling. The referenced data sheet is in the realm of volts of range, on that chip I would expect more like 3v common mode (both outputs) with a 6v power supply input.

The IC (8.1) driving the Hall sensors is run on -6v and +12v, with a pullup diode to +6v, that I suspect is a clamp preventing voltages higher than the 6v supply, thus the hall devices are not being starved of power or current and are being run with the intended supply voltage.

 

[SaSi]

At this point I believe the Hall effect outputs are insufficient to drive the circuit. Either the sensors are damaged, or the magnetic field has diminished. Perhaps the table was dropped and the gap between the ring magnet and the sensors opened up. The DC readings relative to the position of the turntable (I mapped them) were remarkably repeatable, showing the slight difference of flux intensity in the 6 magnets that make up the ring.

The service manual describes a method of re-aligning the hall sensors by means of two screws. I've seen the screws on the motor assembly but it's been a while and don't recall their exact position.

It might be worth to read that section of the service manual and attempt adjustment there.

 

[steerpike2]

It's possible the motor rotor is sitting too high to affect the Hall devices; but this change would have had to happen in the last week, since it worked OK on one phase when this thread was started.

IF you lift the motor rotor out of the bearing sleeve, the machined tolerance is too tight for it to go back in immediately; it will sit on a cusion of air in the bearing-well for days or longer. You could check this. It should have no bounce or compliance when pushed down hard.

Also check the two air-gap screws next to each hall sensor. The gap between the screw tip and the Hall Ic should be about 1mm - but unless it looks SIGNIFICANTLY different (someone played with it???) don't adjust it. It is not an easy asjustment to get perfect after change.

For *these* test you can remove the motor from the sprung upper plinth, and operate it with no platter (platter is only needed for correct testing of servo: with no platter it rotates in a very stepped manner, but that's ok for now)

 

[Robie]

I think steerpike2 is right. I don't think a fall or the motor rotor removal/lube is the problem. The platter did rotate on one phase previously after the rotor removal and maintained speed per the LED speed indicator. Furthermore, I know it hasn't taken a fall while in my possession and was packed in its original box for at least 10 years according to the former owner. No idea as to what treatment it may have received previously while packed away or before that.

I can say definitively (and not defensively), that I never touched any of the screws on the motor adjustment screws or otherwise.

I am humbled by Mark's efforts thus far even though I'm not sure I completely understand what they reveal. I certainly appreciate his perseverance.

However, given my lack of comprehension, in the vein of "just throwing parts at it" (they're cheap regardless) and based on what I was told by JM Technical (the Revox Service Center in the US), don't the ICs have to be ruled out first? Since there is now one or more substitute LM358 ICs installed in the motor control circuit in lieu of the specified ICs (R4559), wouldn't it make sense to replace all the ICs on the motor control board (R4559s and the larger ICs) so they can be definitively ruled out as the problem?

 

[markthefixer]

That's the whole ideas of the sockets

I have just lifted the platter on and off, and if the magnet can be lifted away from the sensors it would certainly cause the reduction in magnetic signal strength I have seen at the sensors.

But I have a phlebotomist appointment at 1:50 and must get my snowbound car out on the street and ready to go, so I'll be back on after that, 3 pm maybe...

 

[steerpike2]

don't the ICs have to be ruled out first? Since there is now one or more substitute LM358 ICs installed in the motor control circuit in lieu of the specified ICs (R4559), wouldn't it make sense to replace all the ICs on the motor control board (R4559s and the larger ICs) so they can be definitively ruled out as the problem?

From all the descriptions given so far about its symptoms, there is no way any of the ICs - other than IC10 - can cause this behaviour.
+6V on the orange wire of the motor is the only signal the motor gets from the servo electronics. If 6V is present, that's all there is to it - the servo is doing as much as it can.

Everything is pointing to the Hall devices - or some mechanical condition associated with their correct operation. The strangely curious part is WHAT happened to them - since they dont seem to have been mistreated in any way. COULD the clamping diode that limits their drive voltage be intermittently open, or have a bad solder joint, or cracked pcb trace? That would put +!2V on the Hall devices and damage them.

I haven't looked at an LM358 data sheet, but I'll trust your collective judgement that it *is* a suitable drop in for the RC4558P. (Or 4559)

 

[Robie]

OK, thanks Mark. Let me know when you are ready to have me order them. Sorry if I appear impatient. Thanks for your efforts.

 

[markthefixer]

I'm back, and do not anticipate leaving the house before next Wednesday. So if you want to drop on by (from now onwards) to either pick it up, or spend a little time with it, I'm game.

I want to see what you did about taking things off the motor, and repeat them with an enlightened, gimlet eye watching. As well as instrumentation....

Garsh, that HP 5.5 digit DMM was one of the best investments I ever made...

 

[steerpike2]

I'm not sure if ReVox used Siemens, but I've discovered the Siemens Hall Ics of the 1970s were made from ferrite - the actual chip body that is. That means they are physically quite fragile. Could they have been physically struck by anything and cracked?

 

[Robie]

I'm not sure if ReVox used Siemens, but I've discovered the Siemens Hall Ics of the 1970s were made from ferrite - the actual chip body that is. That means they are physically quite fragile. Could they have been physically struck by anything and cracked?

Not while the TT was in my possession.

 

[markthefixer]

Something doesn't add up. The two hall sensors are in series, with about 6 volts across BOTH of them. One sensor gets 6.12v to 6.07v on it's power pins, 50 millivolts... the other in series gets 6.07v to 0, 6.07 volts.... No load / voltage sharing etc. Both output pins of one sit near the maximum voltage at 6.09v, while supposedly the internal circuitry of one work at less than a tenth of a volt, while the other has 6.07 volts across the power pins and both output pins sit near 6.06 volts.

And yet BOTH of the $%$# things have a position dependent output that is magnetically coupled. And repeatable.

I would love to get the readings of another turntable's J3 pins 2,3,4,5 and 8 ground referenced (J3 pin 7) voltage. To see if this is a normal or aberrant situation with the hall whatever the h#** they are's.

The motor top plate has a honking BIG thick ring magnet, then a pc board, with the hall effect sensors facing away from the ring magnet, then a metal plate (square) with two screws that concentrate the magnetic flux and approach the faces of the hall sensors. ALL they set is the gap, and thus the strength of the signal coming out of the sensors.

 

[SaSi]

Something doesn't add up. The two hall sensors are in series, with about 6 volts across BOTH of them. One sensor gets 6.12v to 6.07v on it's power pins, 50 millivolts... the other in series gets 6.07v to 0, 6.07 volts.... No load / voltage sharing etc. Both output pins of one sit near the maximum voltage at 6.09v, while supposedly the internal circuitry of one work at less than a tenth of a volt, while the other has 6.07 volts across the power pins and both output pins sit near 6.06 volts.

And yet BOTH of the $%$# things have a position dependent output that is magnetically coupled. And repeatable.

I would love to get the readings of another turntable's J3 pins 2,3,4,5 and 8 ground referenced (J3 pin 7) voltage. To see if this is a normal or aberrant situation with the hall whatever the h#** they are's.

The motor top plate has a honking BIG thick ring magnet, then a pc board, with the hall effect sensors facing away from the ring magnet, then a metal plate (square) with two screws that concentrate the magnetic flux and approach the faces of the hall sensors. ALL they set is the gap, and thus the strength of the signal coming out of the sensors.

OK, I will pull out my B795 and B791 and do a measurement. The B791 locks speed in both 33 and 45rpm (but sometimes it takes a couple of minutes when started after receiving no use for days). The B795 doesn't lock speed at all but visually (with a strobe disk) speed is stable. I understand that the 791 is different in terms of motor control board (also has pich control and readout) but I'm fairly certain that the circuit portion under suspicion is identical. The motors are identical.

I understand which pins to check voltages. What isn't clear is if you need AC or DC voltages and if you want these voltages with platter rotating or not.

Please clarify this and I will do the measurements.

PS. I really hope this TT gets succesfully back into service. Mount a good cart and it will really sing.

 

[steerpike2]

Something doesn't add up. The two hall sensors are in series, with about 6 volts across BOTH of them. One sensor gets 6.12v to 6.07v on it's power pins, 50 millivolts... the other in series gets 6.07v to 0, 6.07 volts.... No load / voltage sharing etc. Both output pins of one sit near the maximum voltage at 6.09v, while supposedly the internal circuitry of one work at less than a tenth of a volt, while the other has 6.07 volts across the power pins and both output pins sit near 6.06 volts.

Sounds like the one with 6.07V across its power pins has gone near open-circuit. With no current flowing through them, neither will produce an appreciable voltage drop.

And yet BOTH of the $%$# things have a position dependent output that is magnetically coupled. And repeatable.

They act as current multipliers; so the output is dependent on the supply current. If the supply current is very low, the output will be correspondingly low.

Can you open-circuit the orange wire & check current flow there?

You might also try applying a light presure to them with a toothpick. If one has cracked and has an internal open circuit, a little physical stress on the packaging may get it to operate correctly for a moment.

You could try bridging the power pins of each Hall device separately with a resistor - to allow normal current to flow through the 'other' one, and check if the non-bridged one is then functioning normally.

 

[Robie]

For me, this thread is a bit like watching brain surgery. While I wouldn't have a comprehensive understanding of all the nuances and details, I'd admire the skills of the doctors, nonetheless. Thanks for your continuing interest/assistance, Dr. steepike2, Dr. SaSi and Dr. Mark! (However, after looking at Mark's avatar again I almost typed, "Paging Dr. Howard, Dr. Fine, Dr. Howard!")

I'm guessing that my problem may be anomalous but hope it contributes to the knowledge base on these TTs.

 

[markthefixer]

Well, I didn't follow directions exactly, and got some interesting results.

A current meter in series with the two sensors (no resistors yet) showed 0.23 mA (230 microamps). This is with the rotor off. Rotor on made no difference.(no surprise but still checking). 5.912v / 0.23mA = "25.7k"

Then each sensor was bridged with a 1k ohm resistor , making a voltage divider that would have delivered about 3v each assuming no current, and different voltages depending upon the current drawn, thus making it possible to calculate currents by the balance between the two.
(edit: EXCEPT that I ran out of meters (3, two were watching IC 10 outputs for Hall sensor response strength, edit edit - I just remembered the HP 5.5 digit.... drat...) and couldn't monitor current real time - I THINK it was still 5.6mA before I closed the break and started measuring voltages - it may be written upon a piece of paper I don't have)

I didn't want to fool around with conflicting results as I bridged one, than the other, if they had interacted against each other. In retrospect, with a bit more time, with the setup I did, I could have done this easily. oh well....

The "top" (50mv) sensor/1k resistor combination had 1.222v across it.
5.6mA - (1.222v / 1000 ohms = 1.2mA) = 4.4mA through the Hall element ("278 ohms")

The bottom (6.07v) sensor/1k resistor combination had 4.69v across it.
5.6mA - (4.69v / 1000 ohms = 4.69mA) = 0.91 mA through the hall element ("5.1k ohms ")

When this was done with the rotor in place (but no platter) there was movement of the rotor. Resistance to motion was felt in the rotor. Spinning the rotor yielded a slight squealing noise that I associated with oscillation of the circuit pulsing the drive coils wildly, as the digital meter readings went nuts, even to strobing the hall element power. Putting the platter in place made no difference, the noise was faint and intermittent, with cycling power necessary to settle things down. Sort of felt like a dry bearing - but it disappeared with power off.

BUT, this meant that there was enough output from the Hall elements (maybe from both - hard to say) to get the motor driver transistors conducting -

This oscillation could have been from the setup, I had desoldered the orange wire carrying the hall power from the motor PCB, and added two stranded wires (current measurement)about 1 foot long to make things accessible outside of the TT, two more 1 foot wires were added, one a ground tap at the motor, and the other a tap to the trace between the hall elements series power connection. This tap is the common point between the two 1k resistors, and the TT motor ground was the lower "R2" to ground of the voltage divider, while the current meter connections were closed and the meter removed, and this served as the connection to the voltage divider's "R1".

Of course this was done as to not interfere with the rotor when installed.

The TT is now back in Robie's possession, thus no more readings from me.

Bottom line - one Hall element ("bottom" J3-4(violet), 5(blue), IC10.1 pin 1) is highly suspect.
IC10.2 pin 7, (serving the OTHER Hall sensor) was the zapped op amp element in the original rc4559 chip

and NO, I was not able to eyeball either Hall element, the construction is (top to bottom):

Rotor with honking HUGE ring magnet, 1/2 inch thick, 1 inch wide, 6 inch diameter (massive platter sits on this)

Printed circuit board, foil up, with tach coil pattern, 6 motor coil connections and the leads of the hall elements, which face down

Large metal plate (flux collector?), with security sealed screws (flux concentrators?)facing up, to apparent hall element positions.

 

[steerpike2]

This is the diskette motor from a circa 1990 5¼" 1.2MB floppy disk drive, as present on every home and office PC sold around that time. I doubt any PC scrapyard would even charge money for one now.

Take off the rotor, an LO AND BEHOLD... not two, but three linear Hall devices!

 

[Robie]

Based on Mark's last post and my discussions with him, it appears that one or more Hall sensors are bad. I assume that obviates the need the replace all 11 ICs on the motor control board, even though they all under $1 each.

Is that a correct assumption?

 

[steerpike2]

I assume that obviates the need the replace all 11 ICs on the motor control board, even though they all under $1 each.
Is that a correct assumption?

Yes, correct.
If the 33/45 LED lights up when you rotate the platter by hand at exactly 33 or 45 (the correct speed of which you can confirm with a conventional printed strobe disc) THEN the servo electronics and PLL are working correctly.

There is only one other IC that this test doesnt check (IC8, an RC4559), but you have confirmed its operation by the presence of +6V on the orange motor wire when below correct speed, and 0V when above correct speed.

 

[Robie]

Although I knew this to be the case, I was hoping we were missing something.

Regardless, this whole experience has really reaffirmed the quality of the members hanging around AK. Thanks all for your help and have a happy holiday season!

 

[SaSi]

Sorry to hear the repair wasn't succesful in the first attempt. The idea is not to give up. Will you be looking for hall sensors? Perhaps from a floppy drive?