600 Ohm inputs

[SmokeyJoe]

I've used truckloads of 553x to drive 600 ohm lines.

I made one headphone amp with a discrete push-pull stage hanging off a TL07x. It could comfortably drive loudspeakers until the 1A wall wart opens it's thermal fuse. I ended up adding a R in series with the charging rectifier in the PS so the rails would sag under heavy use and protect the wall wart.

JR

John, is my circuit in any way similar to the 553x designs you mention here?

 

[JohnRoberts]

John, is my circuit in any way similar to the 553x designs you mention here?

I have done lots of different product and all are not the same. I vaguely recall some using the typical cross connected NF balanced output. That looks similar but I won't swear to accuracy.

JR

 

[abbey road d enfer]

Thanks for the suggestion. Excellent results with 10 ohm - doesn't seem to be any roll off at all.

OK. Mystery unveiled.

(As for next steps - is there any benefit to finding a trade off between increasing the output resistor value and acceptable low end roll off?

I don't think tehre is any benefit from increasing teh value of the resistors there. OTOH, the circuit the secondary is connected to may prefer seeing a larger impedance, particularly a passive EQ. You need to chck that. Then this could be addressed by adding resistors between the and the subsequent circuit.secondary

Or ok just to leave them as 10 ohm?

If 10 ohms is good, leave it so.

I believe increasing my supply to +/-15V would be worthwhile to increase headroom.

It would; not terrribly though, only about 2dB. Don't count on it to make a night and day difference.

Perhaps there is a more suitable transformer to use on the input (further research required...)

Certainly? However, you have just experimented with the fact taht it is often possible to alter a circuit for making it work with different parts than originally planned.

Edit: Preliminary tests are showing quite a bit of THD with the driver and Pultec in the loop (~50dB below signal level). The driver alone is very clean, and the Pultec alone is even worse. I'll try and conduct some proper measurements.

You may have to resort to other opamps...

 

[SmokeyJoe]

Thanks for the assistance so far folks.

I have done lots of different product and all are not the same. I vaguely recall some using the typical cross connected NF balanced output. That looks similar but I won't swear to accuracy.

JR

Fair enough! I found it reassuring that you've had acceptable results using 553x opamps - hopefully I can achieve something suitable too.

If 10 ohms is good, leave it so.

That sounds reasonable.

the circuit the secondary is connected to may prefer seeing a larger impedance, particularly a passive EQ. You need to chck that. Then this could be addressed by adding resistors between the and the subsequent circuit.secondary

Noted. I'll see what I can read around it - certainly this circuit has been built in this configuration with the same transformer by a lot of people, so hopefully there is some info out there.

You may have to resort to other opamps...

I made some THD measurements - looks like the driver alone isn't too bad, but once the Pultec is added in the THD increases considerably. I'm not sure what to make of it in all honesty - perhaps different opamps are required, as you suggest.

Worth noting that the opamps are currently adding 6dB gain, so I'll need to tweak some resistor values to get unity (I guess this will also reduce the THD).

 

[abbey road d enfer]

I made some THD measurements - looks like the driver alone isn't too bad, but once the Pultec is added in the THD increases considerably. I'm not sure what to make of it in all honesty - perhaps different opamps are required, as you suggest.

View attachment 111568View attachment 111569

When you say "driver only", is it the driver with the xfmr?
And "driver and Pultec", do you measure the input or the output of the Pultec?
What happens if you load the secondary of the xfmr with a 600 ohms resistor?

 

[SmokeyJoe]

When you say "driver only", is it the driver with the xfmr?
And "driver and Pultec", do you measure the input or the output of the Pultec?
What happens if you load the secondary of the xfmr with a 600 ohms resistor?

Yes driver only is the driver and transformer.

I’m measuring the output of the Pultec (or output of the driver when driver only).

I’ll see what happens with the 600 ohm resistor - although I believe this is not advised with the G-Pultec due to how it interacts with the filter circuitry.

 

[abbey road d enfer]

Yes driver only is the driver and transformer.

I’m measuring the output of the Pultec (or output of the driver when driver only).

OK, so in order to identify the problem clearly, you need two measurements:

• xfmr output unloaded
• xfmr output loaded with 600-620r

I’ll see what happens with the 600 ohm resistor - although I believe this is not advised with the G-Pultec due to how it interacts with the filter circuitry.

A true Pultec has a 620r resistor across its input, which, driven by a 600r source, is supposed to result in nominal frequency response.
However, it would be instructive to assess the real influence of the source impedance over the operation.
Of course level will change, but I think the frequency response and EQ range will not change much between a 600 ohm source and a near-zero one.
At least that's what simulations lead to conclude.

 

[SmokeyJoe]

OK, so in order to identify the problem clearly, you need two measurements:

• xfmr output unloaded
• xfmr output loaded with 600-620r

Looking at the schematic I can see I currently have a 10k resistor across the secondary. So to be clear, I should remove this 10k, and take THD measurements at the output of the Pultec with:

- no resistor across the input transformer secondary (open)

- a 620R across the secondary

And I expect it would also be helpful to make some frequency measurements with these resistors, with the controls in arbitrary positions?

 

[abbey road d enfer]

And I expect it would also be helpful to make some frequency measurements with these resistors, with the controls in arbitrary positions?

Yes. I guess the differences are not enormous, but certainly comparing with the published graphs would give a clue.
Actually I'm 99% certain that whatever the difference, it is usable in practice.

 

[SmokeyJoe]

Yes. I guess the differences are not enormous, but certainly comparing with the published graphs would give a clue.
Actually I'm 99% certain that whatever the difference, it is usable in practice.

Should I be making these measurements with or without the driver?

 

[FIX]

I actually think there lying and it's the same 1k input as the mic. But now we're talking conspiracy.

A proper mic input is 1500 ohms, which is 10x the output of the mic, in most cases they are 150 ohms. Most opamps will drive that fine. If you use the mic input, expect that, if you use the 1/4” input it’s higher so it won’t load a guitar.

 

[FIX]

Should I be making these measurements with or without the driver?

As long as you haven’t been drinking you can drive with out paying for a driver.

 

[abbey road d enfer]

Should I be making these measurements with or without the driver?

you could make it both ways, but only if your source has very low Z. In doubt, do it with the driver.

 

[SmokeyJoe]

you could make it both ways, but only if your source has very low Z. In doubt, do it with the driver.

Alright cheers. Away for a few days but I’ll post results next week

As long as you haven’t been drinking you can drive with out paying for a driver.

I make all my measurements in the car - it has the best sound system! Harder to focus on the road though.

 

[Cranehazard]

A proper mic input is 1500 ohms, which is 10x the output of the mic, in most cases they are 150 ohms. Most opamps will drive that fine. If you use the mic input, expect that, if you use the 1/4” input it’s higher so it won’t load a guitar.

Balanced 1/4 line inputs is what I was referring to. But thanks for your high impedance input.

 

[Cranehazard]

+24 dBu is 17.36 V peak, makes 28.9 mA peak. Which means a push-pull stage needs 14.5 mA quiescent current.

An AD797 happily drives 200 Ohm, with very low distortion. Several composite opamp topologies exist which produce even better results. No need for discrete...

Samuel

Another op amp that can drive 600ohms imagine that!

 

[Cranehazard]

Are you serious or joking? Let's say you have a 6kW contract with the electricity company. You plug a 9kW range/oven. What happens?

In practice there is very little difference with audio. What you are talking is not nearly the same thing. Do you record stuff. Or just operate in the hypothetical. Your comparison is ridiculous!

 

[Cranehazard]

What is the purpose of overcompliicating things and discouraging people I'm sensing alot of negativity. Mocking peoples experiences is weird. This is a diy site some people sound like equipment manufacturers guerilla marketing (just give up and buy our thing, we do it better). All their sqirming just shows you can do better at home. Anything you create will be alot better than the crap they are pushing out. Keep building and running your own experiments. Many things people take as gospel are dead wrong. Misinformation runs rampant in audio usually there is financial motivation behind it.

 

[abbey road d enfer]

In practice there is very little difference with audio. What you are talking is not nearly the same thing.

Think it over. When the overcurrent protection operates it does the same as a resettable circuit breaker, limits delivered power.

Do you record stuff. Or just operate in the hypothetical.

Been a SE and designer all my life.

Your comparison is ridiculous!

I see you have a pretty good opinion of your competnces and a rather bad of others. You may find that some members here have a pretty good record of designing, building and using stuff.

 

[abbey road d enfer]

Quote: +24 dBu is 17.36 V peak, makes 28.9 mA peak. Which means a push-pull stage needs 14.5 mA quiescent current.
I very much doubt Samuel Groner wrote that. Probably taken out of context.
A class A stage would actually require a quiescent current of a litlle over half peak current, but most opamps have an output stage that's close to pure class B, with a quiescent current of less than 1mA.