TGSL Experiments

[ApBerg]

Hello Danny,

The cables from my pcb to the front controls ( disc pot / sens pot , etc ) are twisted.
Could this cause troubles ?

Yes ! keep those wire's as short as possible and do not twist them ! Do use only carbon based potentiometers ( no wire inside ) ,also for the pcb pot's.

Groeten,

Appie

[satdaveuk]

pots

Hi Simon and co
Hope all is well
Talking about pots dont know if I mention this before but one thing I did proved if using all plastic carbon pots with the tgsl you dont need a screened housing all that needs to be bonded to earth is the coil plug.
As soon as you replace with metal bodied types you get falsing when touching the controls, proved this by substitution to prove a point, although only did it with the ground balance.

Anyway just thought Id let you know and that im still around following this thread with great interest
All the best

[der_fisherman]

(See below)

Title:- Great Quotations of the 3rd Millenium (1999 to 2999) to remember!!!!

Dear Dennis the Mennis (great name, badly spelled?)

you said:-

The more things are closer watched the more questions are raised .......

How right you are.......sadly! But at least it stops us all here from getting bored!!!

Regards and have a great day your friend

Andy

[der_fisherman]

Hello Danny,

The cables from my pcb to the front controls ( disc pot / sens pot , etc ) are twisted.
Could this cause troubles ?

Yes ! keep those wire's as short as possible and do not twist them ! Do use only carbon based potentiometers ( no wire inside ) ,also for the pcb pot's.

Groeten,

Appie

You may have identified my problems with my IGSL!!!!!

Tnaks.

[simonbaker]

Simon asked for an example to completely separate Tx and Rx coils :

Here is a clip of the results , the first clip is made with an cable similar to Belden 8723 ; 2 twisted pairs , each pair shielded separatly and ground wire .

The second clip is made with a Belkin usb cable .


I've started both clips in a kind of start position :

Tried to 'null' the coils as much as possible ( I had a 2 meter extension cable made to the Rx coil ) from there I removed the Rx coil away about 1,5 - 2 meter from the Tx coil.

The extension cable has 2 twisted pairs but only each wire is shielded .

The separation of Tx and Rx is a great way for noise hunting !

It made made me painfull clear that I was operating on the edge of the noise ( the biggest noise source ) : the two laptops on my desktop .

In the test setup I usually do , I will keep the next time a little more distance from my desktop .

The testpoints in the clip are the output LF353 (red) vs the output of the LM308 (green)

I've added two extra measurements in the bottom :

- the peak to peak values for the sample time at that moment ( I've made serveral)
- the max , min and average (noise) frequency for the sample time

So please set the quality in Youtube in 720p HD and Full Screen for max result.

In the end of the first clip I've tried to move the Rx coil a little closer towards my desktop ,so this explains the little more noise at the end ...

Had some difficulties with triggering in second clip ...


1) Belden 8723 (similar cable) http://youtu.be/hgXPTS5e4FQ?hd=1

2) Usb cable (Belkin): http://youtu.be/yrtVT2oSsaw?hd=1


In this example the Usb cable with my pcb ( the old TGS ) seems like an untamed beast ... I don't know why yet

The cables from my pcb to the front controls ( disc pot / sens pot , etc ) are twisted.
Could this cause troubles ?

The more things are closer watched the more questions are raised .......


kind regards

Danny

Dennis the Mennis

Hi Danny - thanks for the tests, excellent.

I would be interested in a summary of your conclusions from the tests; what do you think they showed?

I will say what I think, but I may not be understanding correctly, since I cannot see the coils while you were testing, and I may not correctly understand the scope display.

It looks like you showed the noise signal at LM308 output averaged over different time intervals with the coils very separated.

For the Belden 8723, the average noise seemed to be around 30 mV pp. The average frequency of the noise seemed to be around 6 Hz. The RX signal seemed around 60 mV pp. I don't know how to figure the frequency of the RX signal -- it seems to have a different time scale that is not shown.

I could not tell if you took a measurement with the coils nulled. I would like to see a comparison with the coils nulled during the same test.

For the USB cable -- both signals (the null and the noise) were much larger.

The noise signal seems around 650 mV pp or more. The RX signal seems around 2300 mV pp.

I'm not sure how to interpret the USB cable test. Could it have been a mistake in the connections? Were the coils not separated as much as for the other test (still wouldn't explain the high noise value)? Could the scope probes have had a problem? Do you have another USB cable you can try?

I find it best to re-try such tests many times. I never trust my experiments when I do them once. I'm sure time is limited and at least you got some good data for us to think about.

Anyway, I hope you can do those tests again, and include a measurement with the coils nulled and then separated far apart (a label on the video would help in understanding which case is which). I agree something looks funny with the USB cable test.

Also, can you explain the time scale of the RX signal?

Good stuff!

-SB

[simonbaker]

Hi Simon and co
Hope all is well
Talking about pots dont know if I mention this before but one thing I did proved if using all plastic carbon pots with the tgsl you dont need a screened housing all that needs to be bonded to earth is the coil plug.
As soon as you replace with metal bodied types you get falsing when touching the controls, proved this by substitution to prove a point, although only did it with the ground balance.

Anyway just thought Id let you know and that im still around following this thread with great interest
All the best

That is very interesting, I would like to know if anyone else can do that same test. I certainly makes some sense. I have always worried about those long wires going to the controls, especially since they connected to critical areas in the RX pre-amp and also the sync pulses. As for twisting the wires, I'm never sure, but that would be fun to test and measure. Perhaps when you twist the wires, the ground wire carries noise from your hand to the PCB ground and couples it to the other wires on the way -- exactly what you don't want!

I have some pots with long plastic sticks for knobs and that can be mounted right on the PCB, which would avoid the wires completely. However, it would cause a gap in the case due to a hole for the stick-knobs to poke through, letting in dust.

BTW - do you get the problem with metal pot even if you put a plastic knob on it and only touch that?

-SB

[simonbaker]

Hello Danny,

The cables from my pcb to the front controls ( disc pot / sens pot , etc ) are twisted.
Could this cause troubles ?

Yes ! keep those wire's as short as possible and do not twist them ! Do use only carbon based potentiometers ( no wire inside ) ,also for the pcb pot's.

Groeten,

Appie

Glad you mentioned that Danny, I'm sure I twisted them on one of my PCBs (probably got too lazy to do it after that). Thanks Groeten for the advice, it would be interesting to actually demonstrate the difference with a test.

I wouldn't think this would have affected Danny's recent tests, unless he happened to have his hand on the controls for the USB test -- perhaps that happened???

-SB

[golfnut]

Yes, P1 and P2 have 20V p to p Tranmit signal on them,

I had to route my wire (twisted) bundles and tack them with glue dots to pcb, to get stability during sweeping.

It may be advantage to take 3 coax's out to pots. Or extend the shafts to pots inside the case.

S

[Dennis the Mennis]

Hi Danny - thanks for the tests, excellent.

I would be interested in a summary of your conclusions from the tests; what do you think they showed?

I will say what I think, but I may not be understanding correctly, since I cannot see the coils while you were testing, and I may not correctly understand the scope display.

It looks like you showed the noise signal at LM308 output averaged over different time intervals with the coils very separated.

For the Belden 8723, the average noise seemed to be around 30 mV pp. The average frequency of the noise seemed to be around 6 Hz. The RX signal seemed around 60 mV pp. I don't know how to figure the frequency of the RX signal -- it seems to have a different time scale that is not shown.

I could not tell if you took a measurement with the coils nulled. I would like to see a comparison with the coils nulled during the same test.

For the USB cable -- both signals (the null and the noise) were much larger.

The noise signal seems around 650 mV pp or more. The RX signal seems around 2300 mV pp.

I'm not sure how to interpret the USB cable test. Could it have been a mistake in the connections? Were the coils not separated as much as for the other test (still wouldn't explain the high noise value)? Could the scope probes have had a problem? Do you have another USB cable you can try?

I find it best to re-try such tests many times. I never trust my experiments when I do them once. I'm sure time is limited and at least you got some good data for us to think about.

Anyway, I hope you can do those tests again, and include a measurement with the coils nulled and then separated far apart (a label on the video would help in understanding which case is which). I agree something looks funny with the USB cable test.

Also, can you explain the time scale of the RX signal?

Good stuff!

-SB

Hi Simon ,

Well my summary would be uhhh ..... .

What I don't understand is this huge difference in cable choise for my TGS pcb and most likely others will or can run into this problem of cable choise ( for TGSL or IGSL ? )

Where others have no problem the usb cable , this could be for my pcb.

So you are right that these tests should repeated more as once

I will say what I think, but I may not be understanding correctly, since I cannot see the coils while you were testing, and I may not correctly understand the scope display.

It looks like you showed the noise signal at LM308 output averaged over different time intervals with the coils very separated.

For the Belden 8723, the average noise seemed to be around 30 mV pp. The average frequency of the noise seemed to be around 6 Hz. The RX signal seemed around 60 mV pp. I don't know how to figure the frequency of the RX signal -- it seems to have a different time scale that is not shown.

Yes you are exactly right Simon , you saw exactly what I perhaps should have explained more detailed .Sometimes it's more clear in your head and hard to bring it back to paper .

Offcourse it is a moment in time what the noise concerns.

Ididn't mean anything with the Rx signal except that it is hard display while showing the noise on the output of the LM308 on the other channel.

It was more to show amplitude of the Rx signal vs the noise (on your request ?? )

I could not tell if you took a measurement with the coils nulled. I would like to see a comparison with the coils nulled during the same test.

Both clips started with the coils 'nulled' ( = tried to get the lowest Rx signal possible givin the circumstances on the output of LF353) ) in the best way possible due the extra 2 meter extension cable on the Rx coil / they kept sliding slowly away.

Ater a couple of seconds I removed the Rx coil as far as possible for max (?) decoupling .

For the USB cable -- both signals (the null and the noise) were much larger.

The noise signal seems around 650 mV pp or more. The RX signal seems around 2300 mV pp.

I'm not sure how to interpret the USB cable test. Could it have been a mistake in the connections? Were the coils not separated as much as for the other test (still wouldn't explain the high noise value)? Could the scope probes have had a problem? Do you have another USB cable you can try?

I will double check the USB cable and the second test was an exact copy of the first only with the other cable.

During this test the probes were not removed from the pcb .

This test I did repeatedly though , but I kept the same ( not making an sense ) results.
Will make the test again asap .

The question of the controls I've asked because I had also some motorboating effect the moment I connected my earplugs to the speaker output during the tests with different cables and shieldings ...
Sometimes the controlwires seem to be a little touchy

( what is happenig overhere , there was a time I did not ran into this misery )

kind regards

Danny

Dennis the Mennis


ps I know it should be Menace but this more a kind of poetic freedom , a phonetic way to write down Menace.

[Dennis the Mennis]

Yes, P1 and P2 have 20V p to p Tranmit signal on them,

I had to route my wire (twisted) bundles and tack them with glue dots to pcb, to get stability during sweeping.

It may be advantage to take 3 coax's out to pots. Or extend the shafts to pots inside the case.

S

So there is something to gain for those having problems with their control(wires) ...

Thnx Steve ...

[Dennis the Mennis]

Hello Danny,

The cables from my pcb to the front controls ( disc pot / sens pot , etc ) are twisted.
Could this cause troubles ?

Yes ! keep those wire's as short as possible and do not twist them ! Do use only carbon based potentiometers ( no wire inside ) ,also for the pcb pot's.

Groeten,

Appie

Hi Ap ,

Almost an Elfstedentocht he ? To bad it didn't went along !

I know you have built a lot of TGS pcb's , so I must investigate my twisted control wires .
Or perhaps glue them down like Steve did ... to be continued


kind regards ,

Danny


Dennis the Mennis

[simonbaker]

Hi Simon ,

Well my summary would be uhhh ..... .

What I don't understand is this huge difference in cable choise for my TGS pcb and most likely others will or can run into this problem of cable choise ( for TGSL or IGSL ? )

Where others have no problem the usb cable , this could be for my pcb.

So you are right that these tests should repeated more as once



Yes you are exactly right Simon , you saw exactly what I perhaps should have explained more detailed .Sometimes it's more clear in your head and hard to bring it back to paper .

Offcourse it is a moment in time what the noise concerns.

Ididn't mean anything with the Rx signal except that it is hard display while showing the noise on the output of the LM308 on the other channel.

It was more to show amplitude of the Rx signal vs the noise (on your request ?? )



Both clips started with the coils 'nulled' ( = tried to get the lowest Rx signal possible givin the circumstances on the output of LF353) ) in the best way possible due the extra 2 meter extension cable on the Rx coil / they kept sliding slowly away.

Ater a couple of seconds I removed the Rx coil as far as possible for max (?) decoupling .



I will double check the USB cable and the second test was an exact copy of the first only with the other cable.

During this test the probes were not removed from the pcb .

This test I did repeatedly though , but I kept the same ( not making an sense ) results.
Will make the test again asap .

The question of the controls I've asked because I had also some motorboating effect the moment I connected my earplugs to the speaker output during the tests with different cables and shieldings ...
Sometimes the controlwires seem to be a little touchy

( what is happenig overhere , there was a time I did not ran into this misery )

kind regards

Danny

Dennis the Mennis


ps I know it should be Menace but this more a kind of poetic freedom , a phonetic way to write down Menace.

Thanks for clarifying. I wasn't able to really see the average noise when coils were nulled, so maybe you can lengthen that measurement for clarity. I guess you would need to hold those coils down somehow, it's difficult with a jig of some kind.

I'm curious -- did you have your scope display a different time scale for the RX signal? I assume yes, but just wondering. And yes, it was useful to see it I think.

I would try a different USB cable just to eliminate some defect question.

Cheers,

-SB

[satdaveuk]

That is very interesting, I would like to know if anyone else can do that same test. I certainly makes some sense. I have always worried about those long wires going to the controls, especially since they connected to critical areas in the RX pre-amp and also the sync pulses. As for twisting the wires, I'm never sure, but that would be fun to test and measure. Perhaps when you twist the wires, the ground wire carries noise from your hand to the PCB ground and couples it to the other wires on the way -- exactly what you don't want!

I have some pots with long plastic sticks for knobs and that can be mounted right on the PCB, which would avoid the wires completely. However, it would cause a gap in the case due to a hole for the stick-knobs to poke through, letting in dust.

BTW - do you get the problem with metal pot even if you put a plastic knob on it and only touch that?

-SB

Yes i did still have the problem with plastic knob fitted simon so changed back to all plastic and problem resolved.
The lead twisting is interesting because ive always felt that my machine isnt as sensitive to metal objects as a few builds on here, and all my leads from each pot are twisted neatly together, after seeing comments on here im going to run them singlely and let you know the results.
And it comes to mind if you look at typical Tesro builds of that era they all looked not tidy but quite a mess far as cableing, so im thinking was this done intentionely, certainly makes sence from the findings here.
Will play with this, if there is a difference of any kind will try coaxal type cables as sugested on here.

Regards

[satdaveuk]

Well untwisted my the pot wires which made no difference at all far as sensitivity but after saying that and after reading through the posts again on here everyone seems to be talking about instability/falsing when sweeping etc.
A tip for that I found in the early days is if your using metal package 2n2907/2n2222a, in the TX oscillator stage change them for plastic just beware of pinout connections because they vary with each manufacture, you will proberly cure your problems as I did, if not its back to the drawing board.

[Dennis the Mennis]

Well untwisted my the pot wires which made no difference at all far as sensitivity but after saying that and after reading through the posts again on here everyone seems to be talking about instability/falsing when sweeping etc.
A tip for that I found in the early days is if your using metal package 2n2907/2n2222a, in the TX oscillator stage change them for plastic just beware of pinout connections because they vary with each manufacture, you will proberly cure your problems as I did, if not its back to the drawing board.

Hi Dave ,

What did you solve exactly with changing the metal for plastic housing ?

kind regards ,

Danny

Dennis the Mennis