A Switched Loop Gain Option

[Mike_New]

Hi Folks,

As a result of Karl mentioning the loop gain feedback adjustment on the Bearing thread, I thought that I should separate the following discussion onto a another thread, just in case it takes off.

For some time now I have been aware of the possibility of adjusting the resistance capacitance network on-the-fly by switching in a range of values as you are playing the record.

My consideration for this was prompted some time ago by an engineer in Melbourne who has one of my Platters and was experimenting with values by independently wiring them into the circuit. He claimed he did achieve different sonic results although not always superior ones. I got to thinking, it would be nice to be able to select one of a chosen set of values on-the-fly.

However this approach does involve the soldering of connecting wires into the PCB and the removal of components. I do not need to explain what this means for those that are not electronic buffs.

My original thinking was to provide a small circuit board, which contained 4-6 reed relays maybe more, (these are very small relays in a glass tube). These relays would be selected from an external switch mounted in a box. Each relay would switch in a different value of R or C. The components whose values can be changed, are C210 R208 and R209

Now these component values have been experimented with some time ago but I do not believe any definitive results were ever achieved. Also it is recommended by others to change these values when using the SL1200 as a cutting lathe.

This circuit board was to be wired into the main PCB and physically close to the existing components.

However on investigating the project further I have come to the conclusion that it would be possible to mount the selector switch directly onto the chassis to the left of the slider. At this point there is sufficient room under the chassis to locate the switch and components without interfering with the molded center body of the TT.
I do not think that having the components this far from the control IC AN6680 will affect the feedback performance.
A high quality audio switch will be required and quality wiring.
This option does however require the drilling of a mounting hole into the top of the chassis which some would prefer not to have.

The above approach and options allow for the immediate sonic results of any given combination of RC and would be far more comparative than having to wait some time (maybe days) before hearing the results from the next chosen set of values, by which time any valid comparison is impossible to make.

This could be another and perhaps the last radical upgrade to the venerable SL which further enhances an excellent TT.

[MartinT]

Mike - do you have some ready-reckoner calculations for the component values for different rotational momentum of the various platters on sale?

[OneyedK]

The above approach and options allow for the immediate sonic results of any given combination of RC and would be far more comparative than having to wait some time (maybe days) before hearing the results from the next chosen set of values, by which time any valid comparison is impossible to make.

It's not easy. The three components you mention form the loop filter.
If you change all three at once, you can get a new valid loop filter, wich is hopefully capable of starting the table.
If you start switching individual components, you'll create a large number of invalid loop filter situations, not worth listening too, if that's even possible.

I understand Martin's question.
We need a way to map the speed variations of the platter to the presented variating control voltage. If we knew this, we would know the limitations and avoid overshoot or undershoot. Next step would be to re-engineer the loop filter.
One problem is, this "map" changes with the weight of the platter.
Maybe I'm overthinking all this and should the motor torque (if correctly specified) and the weight of the platter be enough to start calculating...

I hope I expained myself well, would've been easier to explain in Dutch :s

[Mike_New]

Hi Martin,
No I don't at this time, I think it would require a fair amount of detailed knowledge of the parameters of the ICs for any meaningful calculations.
All I can say is that the values used that I am aware of, range as follows:
C210 0.2 to 0.68uF (or higher)
R208 22K to 220K
R209 97K to 330K (or higher)

My friend in Melb. only changed C210 to 0.68uF and 33K in // with R208
effectively making it 28.7K as above..
He said it did seem to make a difference when using my Platter.

DC in his mods, placed a 33k in // with R208 effectively making it 28.7K
and he increased C210 to 0.69uF by placing a 0.47uF in //
Also R209 was replaced with 100K.

When fitting the Bearing and Base Plate to the SL1200 used as a cutting lathe, I was asked to fit a 150K in // with R209 effectively making it 103K.
This single mod is reccomended by the German supplier of the cutting attachment. and apparently does improve the torque when heavy cutting.

If you study the circuit block diagram, you will see that R209 feeds from the phase control output into the RC network, where it feeds into the OP amp, reducing this value would appear to increase the input signal across the resistive devider network.
C210 would obviously provide the capaitive smoothing /delay to any sudden changes of input to the OP amp.

However here is the caveat: if you study the diagram carefully you will see that firstly the input ECR (pin 17) of the OP amp appears to be earthed, but then it feeds into the motor control chip pin 5 as signal EC. ??
ECR (pin6) from AN6675 is shown as originating from the resistance divider network
R207,R206, R208

I think it was Paul H who first indicated to me that the diagram contained some errors.

What would probably work is for a few people to who may wish to try this, to interchange values they have tried.

[Mike_New]

Hi Karl,
As I was typing the above post to Martin you got in before me with your post. It takes so long for signals to reach the UK from Australia!!
It would seem that you are somewhat conversant with this sort of electronics so you could probably contribute to some practical ideas and solutions.

[MartinT]

DC in his mods, placed a 33k in // with R208 effectively making it 28.7K
and he increased C210 to 0.69uF by placing a 0.47uF in //
Also R209 was replaced with 100K.

All I can tell you is that DC's mod was applied to my deck from the start when I purchased it from him. It was fine for the stock platter but did not work at all well with your heavy platter, giving a warble effect to sustained notes especially at 45rpm. I subsequently cut it out.

I must check R209 against the circuit diagram as that may still be non-standard.

[Mike_New]

Hi Martin,
If you have left R209 as it was modded then the chances are that it will still be set to 100K which is virtually the value chosen for extra torque in the cutting lathe application. see above notes.
In my analysis we need less capacitance delay for a heavier platter, hence removal of the extra 0.47uF would have made a difference.

However I have compared more closely block diagram and the circuit diagram, and there is most certainly a difference. The resistors are connected up differently.
The problem is that both variations could possibly be valid.
Also C210 on the PCB is designated C120 on the track side.
I am in the process of tracing out the actual track circuit on the PCB to see which is the correct version and will translate this to the block diagram.

[OneyedK]

However I have compared more closely block diagram and the circuit diagram, and there is most certainly a difference. The resistors are connected up differently.

I made a comparison between the pcb, the schematic and the block diagram.
Couldn't find differences...
It is best to redraw the schematic in a normal fashion to see how it actually works. If I find the time, I will do that today.

Here's the block diagram with "added value".


This is the schematic diagram.


This is the pcb.

[OneyedK]

To prevent us all messing with the wrong components

[MartinT]

Thanks, Karl. I will check my R209 value, which should be 330k, tonight.