Proper wiring to avoid ground loops

[abbey road d enfer]

The termination resistor in these Ampex machines is there only to make sure the VU-meter calibration is correct.
The output stage uses a negative feedback arrangement that compensates for any non-linearities in the transformer. However, it does not compensate for the resistive loss due to the secondary DCR. That's why the actual load must be close to 600 ohms.
VU-meter calibration is important because it's used for the machine alignment, but the less-than-1dB difference in level between loaded and unloaded is negligible in operation.

 

[Brian Roth]

I shoulda said "AG-440" which, along with the ATR-100 machines, were NOT "old tube units".....but solid-state.

Input impedance on an AG-440 depended upon which (if any) accessory transformer was inserted into the rear of each channel. IIRC, the ATR-100 had an internal bridging xfmr. that presented a 50K input impedance "into the world".

Bri

 

[moamps]

Strange wording in these period documents.
I came to know that an "unloaded input transformer" means one which secondary is directly connected to the grid of the input tube, without grid resistor (or one of very high value), so the reflected impedance at the primary is mainly governed by the primary inductnace at LF and by the reflected Miller capacitance at HF. With a pentode as input tube, Miller capacitance is quite small (only a few pF), which reflects as only about 1nF, resulting in a HF -3dB point above 50kHz.
All this results in an input impedance that doesn't go below 2 kohm in the audio range.
Clearly, the Ancients knew the benefits of low-loading mics, low load being high impedance.

Nicely said and concisely explained. The designs from that time should be put into perspective when each dB of amplification was far more expensive than it is today and it was important to squeeze every dB possible. The Telefunken V76 preamplifier with a similar solution is also interesting, which is still very expensive today and a highly regarded preamplifier. Its input transformer is quite complicated and, as far as I've read, quite sensitive to phantom power and DC in general.

 

[drtechno]

The termination resistor in these Ampex machines is there only to make sure the VU-meter calibration is correct.
The output stage uses a negative feedback arrangement that compensates for any non-linearities in the transformer. However, it does not compensate for the resistive loss due to the secondary DCR. That's why the actual load must be close to 600 ohms.
VU-meter calibration is important because it's used for the machine alignment, but the less-than-1dB difference in level between loaded and unloaded is negligible in operation.

If I remembered correctly from transformer couple theory class, you load to keep voltage spikes from counter emf feeding back to the output device as well as stabilize the transmission line's impedance and provides a stable load for the driving circuit. A lot of the operators of 600 ohm gear are not aware of this especially in commercial settings, but gear made with this transformer convention always have this termination. Even if they only put it in the secondary of the input transformer. Its still there.

 

[drtechno]

With very few exceptions, mic manufacturers don't publish the effects of loading on their mics, but these effects are real, and significant enough to be heard even by untrained listeners, with a large majority voting for the highset impedance load..

The way I remember it, a passive device is a higher impedance seeking device because its a voltage source that propagates across the input voltage potential. An active device is a lower impedance seeking device because its a current drive source is seeking a load.

Back to the subject of these unprotected grid circuits. Back then when they made them, there wasn't good choices for grid stops so your small signal pentodes went unprotected due to lack of quality parts. Something like these tubes would have something like a 85-90K resistor. Which I would use one of the modern low noise resistors i use for tube gear, like a Dale CMF60 resistor. https://www.vishay.com/docs/31018/cmfind.pdf

But there is also other things people are not aware of too, with these mics, as the tube heater tied to the cathode to provide a constant voltage source for biasing. Which it also changes some other tube parameters that is used to calculate the circuit.

 

[radardoug]

correct, and you suppose to use them to load the output transformers and the vu meter correctly on those old tube units. But you also need it if you are going into something not transformer input (transformer inputs would have its own termination by default).

Neither the Ampex 440 or the ATR100 are tube units.

 

[drtechno]

Neither the Ampex 440 or the ATR100 are tube units.

Ok well let me examine why.
Btw, this is standard transformer coupled theory and balanced audio theory I learned in school. Interesting that little piece fell to the way side.
So I'm expecting a transformer coupled unit.

EDIT: Yes.440 below.

The ATR 100 is harder to find its real schematics

But looking at this they are using the termination resistance the same way. Because balanced audio is terminated at the input device and this unit is supplied with a termination resistor to terminate the audio output device when the standard affair is not used. Also a lot of newer stuff made dont have this termination and at time you hear people having issues because doesn't have a switchable termination.And a lot of times they fix it by accident with an attenuator that provides the load.

 

[moamps]

Back to the subject of these unprotected grid circuits. Back then when they made them, there wasn't good choices for grid stops so your small signal pentodes went unprotected due to lack of quality parts. Something like these tubes would have something like a 85-90K resistor. Which I would use one of the modern low noise resistors i use for tube gear, like a Dale CMF60 resistor. https://www.vishay.com/docs/31018/cmfind.pdf

I've built quite a few tube preamps and amps and I only used grid stopper resistors when I had problems with oscillations. Most often it was a carbon composite resistor of few kohms and most often it was when I used high gm tubes designed to work on UHF.
In this sense, I would like to ask you for a more detailed explanation of why you think that in all cases the grid inputs should be protected with large resistors. What do we protect tubes with grid resistors from? Thanks.

 

[drtechno]

I've built quite a few tube preamps and amps and I only used grid stopper resistors when I had problems with oscillations. Most often it was a carbon composite resistor of few kohms and most often it was when I used high gm tubes designed to work on UHF.
In this sense, I would like to ask you for a more detailed explanation of why you think that in all cases the grid inputs should be protected with large resistors. What do we protect tubes with grid resistors from? Thanks.

Gridstop is a calculation and not a guess. But I'll save this for another thread because I will have to go into vacuum tube building theory and the anatomy of the tube grid.
I'll give you a short answer till then, a recording studio is not RFI friendly anymore. How much computer junk switching power supply hash in the air, cell towers, how much rfi noise happens now, quite a bit compare to yester years. That is why those designs are obsolete.

 

[ccaudle]

Back to the subject of these unprotected grid circuits. Back then when they made them, there wasn't good choices for grid stops...

Unloaded transformer does not refer to the absence of grid stop resistors, which are in series with the grid (which is a high impedance node). Unloaded transformer implies that there is no grid resistor (not grid stopper resistor), i.e. grid to 0V like you have when using cathode biasing. The transformer would provide the connection to DC reference, but would be a much higher impedance at audio frequencies because of the winding inductance and the reflected primary side impedance. The impedance reflected to the primary would be the grid impedance, which is very high as long as the signal level is not high enough to forward bias the grid.

 

[drtechno]

Unloaded transformer does not refer to the absence of grid stop resistors, .

correct, you were the one that referred to an unloaded transformer having no grid stop. I refer to an unloaded transformer as the absence of a resistance load across the primary or secondary of the input transformer. Because its common practice for passive source devices to bridge into the input circuit and a line output device to be loaded then coupled into the stage.

 

[ccaudle]

you were the one that referred to an unloaded transformer having no grid stop

My misunderstanding then. I was attempting to reply to the second paragraph of post 215. I thought the sentence containing "...I call an unprotected input. As the grid stop is used to prevent excessive current flow...." was still discussing an unloaded transformer, I did not realize that the conversation had shifted to a related topic.

 

[drtechno]

My misunderstanding then. I was attempting to reply to the second paragraph of post 215. I thought the sentence containing "...I call an unprotected input. As the grid stop is used to prevent excessive current flow...." was still discussing an unloaded transformer, I did not realize that the conversation had shifted to a related topic.

I was going to start a tread on the grid stop and the anatomy of the grid in the vacuum tube. But I'm in the process of finding and scanning and de-yellowing it so you can see some of the illustrations. Here is one that is about the spiral wound anatomy of the grid and the flux calculations that they consider when setting the mechanical electron focus internally in the vacuum tube. You know, the series resistance of the secondary can suffice a lot of tube gain part of the formula, but the lifespan of it working in the circuit is short compared to the lifespan of the tube.

 

[abbey road d enfer]

Representing a tube grid as a solenoid is simply wrong.
The windings are supported by pillars that short every turn.

 

[moamps]

Representing a tube grid as a solenoid is simply wrong.
The windings are supported by pillars that short every turn.

This model cannot be applied even if the grid were made really in the form of a coil. The reason for this is that in this case the grid current is divided by each turn and a part goes (or comes from) towards the cathode, so at the end of the coil this current is equal to zero. I am happily waiting for the continuation of the discussion, especially the part on how to optimally calculate the grid stop resistor.

 

[abbey road d enfer]

This model cannot be applied even if the grid were made really in the form of a coil. The reason for this is that in this case the grid current is divided by each turn and a part goes (or comes from) towards the cathode, so at the end of the coil this current is equal to zero.

I agree. There could be a philosophical debate about the antenna effect of a solenoid, though.

I am happily waiting for the continuation of the discussion, especially the part on how to optimally calculate the grid stop resistor.

Personally, I vote for the Stetson-Harrisson method, based on the use of multiple written pieces of paper and a hat.

 

[radardoug]

The Stetson-Harrison method! Are you sure its not the Homburg Trilby method??

 

[abbey road d enfer]

The Stetson-Harrison method! Are you sure its not the Homburg Trilby method??

There are many variants, e.g. the Borsalino-Sorrenti in Italy, or the Gapette-Bada in France.