Impedance matching / bridging between tube gear and modern audio interfaces

[mrtnasty]

Hi I am recording with a Warm Audio WA 67 microphone (output impedance = 200 ohms) into an Ampex 601 (input transformer impedance = 200 ohms), the output (600 ohms) is sent into a tube compressor with 600 ohm input impedance, and output impedance is also 600 ohms. This is sent into the line input of my Focusrite Clarett+ 4 pre (66k Ohm input impedance)

During mastering, I send the digitally recorded signals from my Focusrite Clarett+ 4 pre line out (output impedance 68 ohm) back into my tube gear (either the 200 Ohm or 600 ohm input impedance )

I am trying to make sure I am getting my impedances matching / bridging correct.

My questions are:

1) For recording, I am matching impedances all the way through my signal chain, then bridging the impedance going into the audio interface. Is this the correct way to do this, or should I use a transformer and match the 66k ohm input impedance of the audio interface also?

2) For mastering, should I transform the 68 ohm output to match the 200 ohm or 600 ohm input, or is bridging this ok too?

3) Is there anything else I am missing?

Thanks!

 

[ccaudle]

recording with a Warm Audio WA 67 microphone (output impedance = 200 ohms) into an Ampex 601 (input transformer impedance = 200 ohms)

That is a hard load, most microphones are designed with the expectation that the next device has an input of at least 1.5k to 2k Ohms.

output impedance is also 600 ohms. This is sent into the line input of my Focusrite Clarett+ 4 pre (66k Ohm input impedance)

Presumably the Ampex 601 has transformer output, so check to see if there is a specified maximum load impedance for proper frequency response (no peaking).

Focusrite Clarett+ 4 pre line out (output impedance 68 ohm) back into my tube gear (either the 200 Ohm or 600 ohm input impedance )

Most modern gear has trouble driving 600 Ohm input impedance, and 200 Ohm is practically a short circuit. Distortion will be much higher than specified into such a low impedance load. Most modern line outputs are designed with an expectation that they will be driving an input no lower than 5k Ohm, and usually 10k Ohm or higher.

For recording, I am matching impedances all the way through my signal chain

I do not use antique designs personally, so difficult for me to say what the expectations are of any particular gear you have, but matching input impedance to output impedance is rarely done, because it results in 6dB signal attenuation right away.
The Warm Audio page for their WA 67 microphone states the rated load impedance is 1k, so 600 is low relative to their "rated" impedance. It is not clear whether WA uses that term to mean minimum, recommended, or something else.

For mastering, should I transform the 68 ohm output to match the 200 ohm or 600 ohm input, or is bridging this ok too?

For modern solid state outputs you should use 10k input impedance if your tube gear has a way to select that.

Is there anything else I am missing?

Matching impedance is important when you are driving from the studio to the radio transmitter over analog lines several miles long. For studio use matching output and input impedance is almost never the correct choice.

 

[mrtnasty]

That is a hard load, most microphones are designed with the expectation that the next device has an input of at least 1.5k to 2k Ohms.



Presumably the Ampex 601 has transformer output, so check to see if there is a specified maximum load impedance for proper frequency response (no peaking).



Most modern gear has trouble driving 600 Ohm input impedance, and 200 Ohm is practically a short circuit. Distortion will be much higher than specified into such a low impedance load. Most modern line outputs are designed with an expectation that they will be driving an input no lower than 5k Ohm, and usually 10k Ohm or higher.



I do not use antique designs personally, so difficult for me to say what the expectations are of any particular gear you have, but matching input impedance to output impedance is rarely done, because it results in 6dB signal attenuation right away.
The Warm Audio page for their WA 67 microphone states the rated load impedance is 1k, so 600 is low relative to their "rated" impedance. It is not clear whether WA uses that term to mean minimum, recommended, or something else.



For modern solid state outputs you should use 10k input impedance if your tube gear has a way to select that.



Matching impedance is important when you are driving from the studio to the radio transmitter over analog lines several miles long. For studio use matching output and input impedance is almost never the correct choice.

Thanks for the detailed reply! That makes sense that impedance bridging is the modern standard. But I am still a little confused, because I know tube preamps from the 40s and 50s were designed be impedance matched (both input and output), and if that is wrong, I am wondering why they sound so good that way? Also reading the manuals of these vintage preamps, all the frequency response curves shown are assuming you are matching impedances and I'm afraid the actual frequency response won't be the same if I don't follow the manual.

 

[scott2000]

This has some more insights...Can skip the unity gain part...or not..

https://www.ranecommercial.com/kb_article.php?article=2112

 

[ccaudle]

I know tube preamps from the 40s and 50s were designed be impedance matched (both input and output)

I don't know the turns ratio of the input transformer on a 601, but the first stage on the mic input has a 2.2M grid resistor. It seems unlikely that the input impedance is actually 600 Ohms, because that would imply that the input transformer was something like a 1:57 turns ratio. That is too high to be practical.
Have you actually measured the input impedance?

But to your original point, probably through the 50's that was true, all equipment was transformer coupled tube designs, so there was not much disadvantage to designing low input impedance with transformers on outputs and inputs.

The problem is then how to interface most effectively with gear designed decades later using a different approach. Definitely connecting a Focusrite directly to a 200 Ohm input (or even 600 Ohm) is not going to give the best results. You could make a transformer interface which converts a 200 Ohm input into something like a 5K Oh input. That would be about a 5:1 transformer, should be able to get decent frequency response, but it will drop the output level by 14dB, so you will have to look at the gain structure and what input level that device is optimized for.

When a device with output transformer is connected to the line input of the Focusrite you can check the frequency response, and if there are peaks from undamped resonances you can add a resistor across the transformer output to limit the maximum input impedance seen by that output. Jensen typically had recommendations for a R-C series termination across the output of their transformers, so worth checking if there is data on the particular transformers you have in your gear.

 

[abbey road d enfer]

because I know tube preamps from the 40s and 50s were designed be impedance matched (both input and output),

Not so much regarding microphone connections.
If you check the RCA microphone brochures, you will see that the recommended load for microphones was a transformer with an unloaded secondary, which means quite a high inpedance in the 5-10k ohms range.
Regarding line level connections, some vintage equipment requires to be loaded with the proper termination for linearization of frequency response.
In most cases, inputs should be insensitive to the source they are presented.

 

[mrtnasty]

I don't know the turns ratio of the input transformer on a 601, but the first stage on the mic input has a 2.2M grid resistor. It seems unlikely that the input impedance is actually 600 Ohms, because that would imply that the input transformer was something like a 1:57 turns ratio. That is too high to be practical.
Have you actually measured the input impedance?

But to your original point, probably through the 50's that was true, all equipment was transformer coupled tube designs, so there was not much disadvantage to designing low input impedance with transformers on outputs and inputs.

The problem is then how to interface most effectively with gear designed decades later using a different approach. Definitely connecting a Focusrite directly to a 200 Ohm input (or even 600 Ohm) is not going to give the best results. You could make a transformer interface which converts a 200 Ohm input into something like a 5K Oh input. That would be about a 5:1 transformer, should be able to get decent frequency response, but it will drop the output level by 14dB, so you will have to look at the gain structure and what input level that device is optimized for.

When a device with output transformer is connected to the line input of the Focusrite you can check the frequency response, and if there are peaks from undamped resonances you can add a resistor across the transformer output to limit the maximum input impedance seen by that output. Jensen typically had recommendations for a R-C series termination across the output of their transformers, so worth checking if there is data on the particular transformers you have in your gear.

Im using the 17331 matching transformer that is recommended in the 601 manual for use with low impedance microphones. I think it is about 1:16. I haven't measured the input impedance, how would I do that?

That makes sense to adapt to modern gear using a different approach and use bridging impedance, since that is what they were designed for. Could I do it the other way and transform down my Focusrite to say 5 ohms and leave the input of my at tube gear at 600 ohms? The headphone output of the Focusrite is actually already at 5 ohms.

Trying to wrap my head around putting a resistor on the output of the tube gear before it hits the Focusrite, it was also recommended to me by another person on this forum. So if I put a 600 ohm resistor on the output, the tube amp would always think it's driving a 600 ohm load, and the Focusrite would also be happy because its input impedance is way higher (66k ohms)?

 

[mrtnasty]

Not so much regarding microphone connections.
If you check the RCA microphone brochures, you will see that the recommended load for microphones was a transformer with an unloaded secondary, which means quite a high inpedance in the 5-10k ohms range.
Regarding line level connections, some vintage equipment requires to be loaded with the proper termination for linearization of frequency response.
In most cases, inputs should be insensitive to the source they are presented.

I never understood the RCA manual when it was saying you should have the transformer secondary be unloaded. What defines the input impedance in that case, the tube grid? How do you figure the input impedance is 5-10k in that case?

 

[abbey road d enfer]

Im using the 17331 matching transformer that is recommended in the 601 manual for use with low impedance microphones. I think it is about 1:16. I haven't measured the input impedance, how would I do that?

That makes sense to adapt to modern gear using a different approach and use bridging impedance, since that is what they were designed for. Could I do it the other way and transform down my Focusrite to say 5 ohms and leave the input of my at tube gear at 600 ohms? The headphone output of the Focusrite is actually already at 5 ohms.

On most equipment, the headphone output is generally of lesser quality than the line output. Actually, line outputs are typically presenting an impedance of about 100 ohms, which is find for driving a 600 ohm input. However, they are often not capable of delivering the +24dBu headroom that was some kind of standard at the time.

Trying to wrap my head around putting a resistor on the output of the tube gear before it hits the Focusrite, it was also recommended to me by another person on this forum. So if I put a 600 ohm resistor on the output, the tube amp would always think it's driving a 600 ohm load, and the Focusrite would also be happy because its input impedance is way higher (66k ohms)?

That is absolutely correct. Some equipment have a switch that connects an internal load when driving high-Z inputs.

 

[Brian Roth]

On most equipment, the headphone output is generally of lesser quality than the line output. Actually, line outputs are typically presenting an impedance of about 100 ohms, which is find for driving a 600 ohm input. However, they are often not capable of delivering the +24dBu headroom that was some kind of standard at the time.

That is absolutely correct. Some equipment have a switch that connects an internal load when driving high-Z inputs.

Yep...Ampex AG-440 and ATR-100 tape machines had a rear panel switch labeled TERM which connected a 620 Ohm resistor across the secondary of the output transformer.

This all gets a bit messy in a system with both 600 Ohm terminated inputs/bridging inputs/patchbay.

Bri

 

[abbey road d enfer]

I never understood the RCA manual when it was saying you should have the transformer secondary be unloaded. What defines the input impedance in that case, the tube grid?

The actual impedance is the result of the xfmr's primary inductance, parasitic capacitance, losses and the reflected impedance of the Miller capacitance.

How do you figure the input impedance is 5-10k in that case?

Actually, the resulting input impedance is quite complex and can't be expressed with a single figure. It has peaks and dips. The art of the designer is to make sure that the impedance in the audio range is high enough, typically higher than 1k. The midrange peak can be higher than 10kohms.

 

[mrtnasty]

On most equipment, the headphone output is generally of lesser quality than the line output. Actually, line outputs are typically presenting an impedance of about 100 ohms, which is find for driving a 600 ohm input. However, they are often not capable of delivering the +24dBu headroom that was some kind of standard at the time.

That is absolutely correct. Some equipment have a switch that connects an internal load when driving high-Z inputs.

Thanks! I put a 620 ohm resistor on the output and it sounds similar, maybe a little less harsh, but I still need to do more testing. Also for the output from the focusrite (68 ohms) to the input of my preamp (200 ohms) can I transform down the focusrite to about 5 ohms so that it bridges my preamp? If so, Ive got a line transformer laying around that's 50 ohm primary and 500 ohm secondary, can I use this in reverse? PS I tried headphone output from the focusrite to the input of my preamp and it sounded bad, lofi and lots of noise.

 

[abbey road d enfer]

Also for the output from the focusrite (68 ohms) to the input of my preamp (200 ohms) can I transform down the focusrite to about 5 ohms so that it bridges my preamp? If so, Ive got a line transformer laying around that's 50 ohm primary and 500 ohm secondary, can I use this in reverse?

I think the xfmr will cause more issues than it will solve. It is quite likely that the Focusrite is capable of driving 200 ohms with less distortion than teh transformer.

 

[ccaudle]

It is quite likely that the Focusrite is capable of driving 200 ohms with less distortion

I found a picture of a different model in the same family showing NJM5532 devices, so perhaps so. Max output of 18dBu, so that output current is getting close to the maximum short circuit current for the device. Connecting to a 600 Ohm input would definitely be preferred to 200 Ohms.

 

[abbey road d enfer]

5532 can be used to drive 150 ohms, at the cost of distortion.
Indeed, 200 ohms is a difficult load.

 

[Cqwet Dbdfte]

Hi I am recording with a Warm Audio WA 67 microphone (output impedance = 200 ohms) into an Ampex 601 (input transformer impedance = 200 ohms), the output (600 ohms) is sent into a tube compressor with 600 ohm input impedance, and output impedance is also 600 ohms. This is sent into the line input of my Focusrite Clarett+ 4 pre (66k Ohm input impedance)

During mastering, I send the digitally recorded signals from my Focusrite Clarett+ 4 pre line out (output impedance 68 ohm) back into my tube gear (either the 200 Ohm or 600 ohm input impedance )

I am trying to make sure I am getting my impedances matching / bridging correct.

My questions are:

1) For recording, I am matching impedances all the way through my signal chain, then bridging the impedance going into the audio interface. Is this the correct way to do this, or should I use a transformer and match the 66k ohm input impedance of the audio interface also?

2) For mastering, should I transform the 68 ohm output to match the 200 ohm or 600 ohm input, or is bridging this ok too?

3) Is there anything else I am missing?

Thanks!

Impedance matching is needed to preserve POWER (watts, milliwatts etc), voltage matching not critical. A low impedance source can drive a high impedance input, no problem, but not vice versa.
Low impedance lines in pro audio is to preserve signal integrity, avoid ground loops, etc.
Transformer coupling accomplishes both ground (loop) isolation and impedance transformation.

 

[fazer]

Or rather than a transformer Go with a buffered opamp like
Classic Audio Products, Inc.

Or 2520 style for a balanced input.

Classic Audio Products, Inc.

It will present a balanced 10k input impedance. Look at an API console line input for the circuit

Look in the tech section for the input section. I prefer the sound of the 2520 but the hybrid works great in the circuit and in some ways is more neutral.

Look for the active balanced 2520 app notes on this thread

API

Also the capi summing amp card can be used as a pcb card to build on with some minor modification.

 

[abbey road d enfer]

Impedance matching is needed to preserve POWER

Not really. Half the power is lost in the source impedance. The main reason for Z-matching is either the capacity of driving long lines (hundreds of miles) or the dampîng of unwanted resonances in transformers.

 

[gyraf]

So yes, impedance matching is needed when working with lines longer than the shortest wavelength involved, to control the effect of standing waves (reflections from ends). For 20KHz bandwidth, some 14.39 km (~8.94 miles) is our limit for bridging type transfer. Perhaps enough for most of our needs..

/Jakob E.

 

[Cqwet Dbdfte]

Not really. Half the power is lost in the source impedance. The main reason for Z-matching is either the capacity of driving long lines (hundreds of miles) or the dampîng of unwanted resonances in transformers.

Ha ha!
https://www.analog.com/en/technical...ce-matching-calculations-and-simulations.html

Preservation of power is not that important in low power audio circuits, but critical in RF to also keep noise low, and prevent cables and components to be destroyed.