Heathkit Q meter movement won't zero

[orbanp1]

Originally Posted by regenfreak

...
The sector angle of the Heathkit Q meter scale is annoyingly 100 degrees! Out of luck again. I believe the standard scale sector angle is 90 degrees for most analogue meters I got.
...

Hi regenfreak,

Some ideas:

- Chances are that some of the meters would do 100 degrees, even if it is scaled to only 90 degrees. Sometimes the meter movements have mechanical stops to limit the motion, those stops possibly could be adjusted.
Question is if the meter is still linear at that expanded range. You would have to look at the motion of the moving coil in the magnetic gap, the coil should not leave the range where the flux lines are still regular. You can see that in the first picture of the AVO repair web page.

- May be one could scale down rotationally the image of the face plate from 100 deg to 90 deg.
I was looking briefly around for an app that would do such a transformation, so far no luck.
But, thinking about it, such a transformation is very simple when the object (the face plate image) is described in the polar coordinate system!
It is simple linear scaling in the polar coordinate system. The raster scan image is described in the Cartesian (x-y) coordinate system.
The algorithm would look as follows:
- convert the image to polar coordinate system
- scale it from 100 deg to 90 deg
- transform back from polar coordinate system into Cartesian coordinate system
If you have done any image processing programming, this is very simple.
(I am about 10 years out of writing computer graphics sw.)

- You could just lop-off the first 10 deg of the scale. In any event, a meter movement is quite inaccurate in that region, and then just use the remaining 90 deg, covering the original region from 10 deg to 100 deg.
You need to make sure that the meter movement would not do anything if it gets excitation to be in the original first 10 deg of the region.
You can do that mechanically, like the needle is forced against the zero stop (that represents 10%). You need to have a mechanical stop at zero, and you need to adjust the real zero point of the movement so that it would be "below" that forced zero point. If you think about it, some cars had speedometers like that, where the scale started at 10 kmph, or so.
You can do that electronically, too, like adding a diode in series with the meter movement. Obviously you have engineer the circuit so it would "disable" the meter in the first 10%. (To me, this seems to be the most feasible solution.)

- Or you could do just nothing, you would be still left with a very usable Q-meter ;-)

Again, I am hoping I am not confusing anyone here. But I think I will stop now...

Regards, Peter

[wd40addict]

You could look out for a MM-1U multimeter and swap the meters and scale plates over. Sadly the more common Heath boxes use a 200uA movement so they are no use as donors.

[regenfreak]

Peter
Very interesting ideas!! Thanks!

May be one could scale down rotationally the image of the face plate from 100 deg to 90 deg.
I was looking briefly around for an app that would do such a transformation, so far no luck.

I have done this Photoshop CS in a minute using a tool called Puppet Wrap by put the anchor point at the center of the movement. See attachments
below:

I. before transformation at 100 degrees
II. apply puppet wrap,
III. after transformation at 90 degrees


I have taped a compass on my computer screen and used photos taken with a mobile phone camera for this demonstration, which has lots of wide angle lens distortion and key stoning.

In order to do it properly, I will have to use my professional camera and lens mounted on tripod with the focal plane dead parallel to the meter scale's plane. The downside of Puppet Wrap is that it is a kind of finite element mesh transformation that is non-linear and somewhat aribrary! ( I was doing professional fashion and beauty studio photography for a number of years and I did lots of advanced transformation on fashion models's faces and limps with Photoshop!!!)

It is simple linear scaling in the polar coordinate system. The raster scan image is described in the Cartesian (x-y) coordinate system.
The algorithm would look as follows:
- convert the image to polar coordinate system
- scale it from 100 deg to 90 deg
- transform back from polar coordinate system into Cartesian coordinate system
If you have done any image processing programming, this is very simple.
(I am about 10 years out of writing computer graphics sw.)

I read books on imaging processing when I was doing my PhD (I was an aeronautical engineering graduate applying pattern recognition technique to detect flow structures). I think you referred to rotational transformation matrix:

https://homepages.inf.ed.ac.uk/rbf/HIPR2/rotate.htm

That can be done with MATLAB (which I am too lazy to learn):

https://www.mathworks.com/help/images/ref/imwarp.html

Chances are that some of the meters would do 100 degrees, even if it is scaled to only 90 degrees. Sometimes the meter movements have mechanical stops to limit the motion, those stops possibly could be adjusted.

I guess you meant transplanting the movement of a working 50uA meter into the original Heathkit Q meter. My idea is to abandon the original Q meter completely but use a replacement meter with customised Q-scale print based on the Heathkit original Q scale.

You need to make sure that the meter movement would not do anything if it gets excitation to be in the original first 10 deg of the region.
You can do that mechanically, like the needle is forced against the zero stop (that represents 10%). You need to have a mechanical stop at zero, and you need to adjust the real zero point of the movement so that it would be "below" that forced zero point. If you think about it, some cars had speedometers like that, where the scale started at 10 kmph, or so.
You can do that electronically, too, like adding a diode in series with the meter movement. Obviously you have engineer the circuit so it would "disable" the meter in the first 10%. (To me, this seems to be the most feasible solution.)

I am a bit lost with this one. Anyway my electronic knowledge is still stuck in the 1940s vacuum era and has not significantly advanced to solid state yet!

For the Heathkit Q-scale with 100 degree sector, it is 0.5uA per degree;
For many analogue 50uA meters with 90 degree sector, it is 0.5556uA per degree.

[regenfreak]

You could look out for a MM-1U multimeter and swap the meters and scale plates over. Sadly the more common Heath boxes use a 200uA movement so they are no use as donors.

Yes 200uA movement is no use

[orbanp1]

Hi regenfreak,

Yes, I meant using a replacement meter movement (with the 90dg motion range) with a "customized" copy of the original scale.

I am not familiar with Photoshop, but if it works, that is great.
As for taking a picture of the original scale, I would think that a "simple" scan with good enough resolution would give as good a result as taking a picture with a high-end camera (probably with a lot more hassle than a simple scan?).

I also used MATLAB, though not regularly, we only had a limited number of licenses at work, it was very expensive.
Our institute did not qualify for educational prices even though we had a number of university students working with us.
But here is a web page with "open source" alternatives to MATLAB:
https://opensource.com/alternatives/matlab

Having had a look at the schematics of the QM-1, after all it is not that simple to shift the zero point of the meter electronically, so I think that is not the "most feasible approach".

Similarly, for programming the image transformation of angular scaling sounds simple enough, but it is very easy to get lost in the details too.
So it is not that simple after all!

I guess sleeping on the ideas helps to reevaluate them, and put them into perspective, the ones at first you thought were so great

Regards, Peter

[regenfreak]

Peter
Other than worrying about people coronavirus panic buying and no food on the super market shelves, I have thought about this problem morning. It is not a simple rotational transformation. I will try to explain my idea and hope this won't confuse people below

The Heathkit meter has a linear Q scale except at small Q range from 0 to 50. The linear region has a Q "angular resolution" of 2.5 per degree from 50 to 250. Therefore it is possible to apply a simple linear compression or "mapping" of a 100 degrees sector to a 90 degrees sector, so a compression or reduction factor of 0.9. The errors for the non-linear section are not important for small Q...I think the meter readings for small Q are not that accurate anyway. The new meter scale would have a Q "angular resolution" of 2.25 per degree from the values of 50 to 250.

There is a ruler and protractor tool in Photoshop. See attachments below. I have checked the Puppet Warp algorithm by drawing the Heathkit Q markings with green and red lines, followed by apply the Puppet Warp using the center of the movement, zero and maximum 250 as anchor points. I reckon the angular compression is sort of logarithmic (e.g. decreasing angular compression towards the zero scale mark). Therefore Puppet Warp is not very accurate for this.

It is perfectly doable to draw a new Q scale from scratch using Photoshop, applying a linear angular conversion factor of 0.9 for the Q readings and having acceptable errors for the non-linear Q region from 0 to 50. The new scale will have exact radius or center of rotation of the replacement meter's movement, it is very easy to size print in the Photoshop using ruler tool.

I can get MathLab free at work but I like get my hands dirty doing practical stuff than theory

Having had a look at the schematics of the QM-1, after all it is not that simple to shift the zero point of the meter electronically, so I think that is not the "most feasible approach".

Argh, let me think about this

I would think that a "simple" scan with good enough resolution would give as good a result as taking a picture with a high-end camera (probably with a lot more hassle than a simple scan?).

yes scan would do. I don't have a scanner at home but can use the one at work. It takes no time to capture the photo with my camera, it just takes a few minutes but I am too lazy to take the gears out of storage.

[orbanp1]

Hi regenfreak,

I think we are talking about the same thing!
I never called that a "simple rotational transformation" (that is different, I did a lot of those), but a rotational scaling.
Probably the terminology is not accurate. It is a linear scaling in the rotational coordinate space (polar coordinates).
You call it a "linear angular conversion factor of 0.9". To me that is the same thing!
If Photoshop does this for you, it is great!

I am not really familiar with the capabilities of high-end cameras, but when I took a picture of a page with my cheap camera as I was away from my scanner, I used a tripod and I had to fiddle around quite a bit with the set up. This is possibly a lot easier with professional equipment...

The problem I see is with the electronic solution for shifting the zero point of the meter is that it works with such low levels. It takes signal from a diode detector.
A simple diode would not work, as the forward voltage on it would be comparable to the voltage output from the detector, one would need a more complex circuit.
Probably a lot simpler just to make the scale work.

Any luck with balancing?

Regards, Peter

[regenfreak]

Dear Peter

If Photoshop does this for you, it is great!

If you are a good draftsman, the new scale can be drawn by hand on a technical drawing board. So it is an authentic vintage restoration! That's what I would do if I retire and have lots of time.

Any luck with balancing

This will be the last option as I don't trust my dexterity with shifting tiny balance weight with the tip of a temperature-controlled solder iron tip. I have this mental image of myself cursing and swearing when I mess it up with the movement swinging around.

I am still waiting for the shipment of the replacement caps for the Q meter. Once I power it up, I will see if I can just leave it alone or do something more elaborated. Since the original factory made test coil is missing (which is common), I will still need to send someone who has an accurate Q meter to measure the Q of a homebrew test coil of 250 microhneries or 1mH at two test frequencies for calibration of my Q meter.

[stuarth]

If the degree of unbalance is small, you can add a bit of weight with a spot of something like nail varnish to the appropriate limb of the cross. If is looks like a weight has fallen off (maybe the case with your meter, it looks like the needle is heavier than the opposite end of the cross), you could make a weight with a few turns of fine solder, again held on with nail varnish.

The main problem with the nail varnish solution is that you have to wait for the solvent to evaporate to see how the balance is going, but for me, the wait is better than the risk of touching the assembly with a soldering iron.

If you do have to use a different movement, I think a 90 degree movement used over a 100 degree arc will be fine for your Q meter application.

Stuart

[regenfreak]

Thanks Stuart for the suggestions.

If the degree of unbalance is small, you can add a bit of weight with a spot of something like nail varnish to the appropriate limb of the cross. If is looks like a weight has fallen off (maybe the case with your meter, it looks like the needle is heavier than the opposite end of the cross), you could make a weight with a few turns of fine solder, again held on with nail varnish.

The main problem with the nail varnish solution is that you have to wait for the solvent to evaporate to see how the balance is going, but for me, the wait is better than the risk of touching the assembly with a soldering iron.

This looks less risky. I didn't know nail varnish is sticky

If you do have to use a different movement, I think a 90 degree movement used over a 100 degree arc will be fine for your Q meter application.

The "lazy man's" alternative is to use a replacement 50uA meter, keeping its original 90 degree uA scale. For the majority of the linearity region of Heathkit Q meter, since

Q=250 corresponds to 50uA full scale deflection

One can make a simple look-up conversion table or curve for different values of current in uA on the replacement meter versus "the expected Q" on a faultless Heathkit meter.

For example, a reading of Q = 200 should correspond to the current of 40uA in a fully balanced Q meter assuming linearity (Q is proportional to the current reading of the meter; I think Q is also proportional to the resonance voltage across the tank circuit). So the replacement meter should indicate 36uA (200/250 x 50uA x 0.9=36uA).

Similarly,

multiplier mark X1 corresponds to 21.6uA on the replacement meter

multiplier mark X2 corresponds to 12.6uA on the replacement meter

So a two red marks can be drawn on the replacement meter's scale for X1 and X2 as datums

For the non-linear Q reading below 50, simple interpretation is needed using a protractor to measure the non-linear division markings and calculate the equivalent readings on the replacement 50uA meter.

[regenfreak]

The "lazy man's" alternative is to use a replacement 50uA meter, keeping its original 90 degree uA scale. For the majority of the linearity region of Heathkit Q meter, since

Q=250 corresponds to 50uA full scale deflection

One can make a simple look-up conversion table or curve for different values of current in uA on the replacement meter versus "the expected Q" on a faultless Heathkit meter.

For example, a reading of Q = 200 should correspond to the current of 40uA in a fully balanced Q meter assuming linearity (Q is proportional to the current reading of the meter; I think Q is also proportional to the resonance voltage across the tank circuit). So the replacement meter should indicate 36uA (200/250 x 50uA x 0.9=36uA).

Similarly,

multiplier mark X1 corresponds to 21.6uA on the replacement meter

multiplier mark X2 corresponds to 12.6uA on the replacement meter

So a two red marks can be drawn on the replacement meter's scale for X1 and X2 as datums

For the non-linear Q reading below 50, simple interpretation is needed using a protractor to measure the non-linear division markings and calculate the equivalent readings on the replacement 50uA meter.

I can see faulty assumption in my calculations; they are incorrect. It is actually not as simple as that. It is a mapping of two different circular sectors of scale. Mathematically it is a kind of conformal transformation or mapping that name gives me an instant headache

[regenfreak]

OK I can see all the balance weights, they are three brass micro-size coils exactly like what Peter described above. It seems not just a simple out of balance problem, but more like the fault with the adjustment slot mechanism..I need to study how it "activates" the movement.

I almost broke the movement. I accidentally brushed my finger lightly against the movement bending it badly out of shape and almost snapping it!

I got confused and I should focus on the angle in my previous calculations instead of thinking of current. Then the problem is simpler:

The Heathkit Q scale has 5 main division: 50, 100, 150, 200, 250. Each main division has a sector angle of 20 degrees adding up to 100 degrees. Each main division has 10 subdivision with sector angle of 2 degrees each.

The generic 50uA meter has a 90 degree sector scale. I will draw the new Q scale using Photoshop by dividing it into 5 main sectors of 18 degrees each. I divide the main sectors into 10 subdivision with sector angle of 1.8 degrees each. For the non-linear sector below Q=50, I will measure the angle of each subdivision of the original Q markings and re-scale them fitting into a 18 degree main section.(while maintaining angular proportionality of the non-linear scale).

If I had snapped the movement, I would certainly have to use a replacement meter. It is so fragile!

[orbanp1]

Originally Posted by regenfreak

...
Since the original factory made test coil is missing (which is common), I will still need to send someone who has an accurate Q meter to measure the Q of a homebrew test coil of 250 microhneries or 1mH at two test frequencies for calibration of my Q meter.

Hi regenfreak,

I guess you could try calibrating the Q-meter yourself!

If you "trust?" the level meter in the Q-meter, and can read the frequencies accurately, the Q is equivalent to the quotient of the center frequency and the -3dB bandwidth.
You could verify the level meter (or measure the levels) externally by an RF mV-meter, or with a scope.
Get the frequencies with a frequency counter.
So you need to find the peak, then the frequencies where the level indication is ~0.71 times the peak.
Worth a shot, probably better than nothing.

Regards, Peter

[Ed_Dinning]

Hi Sing, I have a set of advance standard coils that are used to check the cal of Q meters, see if anyone local has something similar

Ed

[regenfreak]

Dear Peter and Ed

Sorry for the late reply. Life is crazy in London with the Covoid-19. I am in social distancing mode.

Ed I don't need the standard test coil but just to post my homebrew coil to anyone who has a very accurate Q meter. All I need is the Q values at two frequencies.

I have managed to replace the two electrolytic caps and powered it up. It does not work; the unit is not responsive to the all the variable gang and dial inputs. So there may be more than one fault down the line that I will have to debug stage by stage. I will check the oscillator first. All the tubes' filaments, oscillator coils,the power supply transformer and variable gangs etc are checked ok. Some carbon resistors are a bit out. Voltage regulator OD3 gives out beautiful purple glow.

The silver plated trimmer cap looks in a bad state that is probably needed to be replaced. I have been cleaning all pots with Deoxit.