Schiit KARA Preamp and Headphone Amp Review

[solderdude]

Someone tell me that there is no issue with this, and why this issue ought not be addressed.

What issue should be addressed ? The output resistance of 75 ohm (SE) or 600 ohm (Balanced) ? Why would that be an issue ?

The headphone out circuit design does not seem to work well when loaded with a low impedance load.
So think 50 ohm and lower, it seems to work fine for higher impedance loads like 300 ohm or higher.

As the pre-amp is the same circuit but without the higher power continuity output stage and will only have to drive a few k ohm as the 'heaviest load' it seems to be no problem for the same circuit.

Of course, in passive mode the output impedance will be higher (vary depending on volume setting) which could limit the max. FR when longer or unusually high capacitance interlinks are used. That's where the 'transparent' buffer circuit comes in though which is low enough in output R to drive about any power amp input.
Every passive controller with a volume control will have this issue. Also LDR and transformer coupled ones.

 

[Noske]

What issue should be addressed ? The output resistance of 75 ohm (SE) or 600 ohm (Balanced) ? Why would that be an issue ?

The headphone out circuit design does not seem to work well when loaded with a low impedance load.
So think 50 ohm and lower, it seems to work fine for higher impedance loads like 300 ohm or higher.

As the pre-amp is the same circuit but without the higher power continuity output stage and will only have to drive a few k ohm as the 'heaviest load' it seems to be no problem for the same circuit.

Of course, in passive mode the output impedance will be higher (vary depending on volume setting) which could limit the max. FR when longer or unusually high capacitance interlinks are used. That's where the 'transparent' buffer circuit comes in though which is low enough in output R to drive about any power amp input.
Every passive controller with a volume control will have this issue. Also LDR and transformer coupled ones.

All I asked is a test to measure how output ohms varies with frequency. on the preamp in active mode. Any discussion related to the headphone or passive mode is irrelevant.

Focus.

I quote -

"I had high hopes that the same preamp would be driving the headphone output and in differential mode. Neither is the case. It seems a separate design is carrying that duty which can't keep up with the performance of the preamp"

This has been done with Saga in active mode, the link of which I have provided, and a quote provided by our generous benefactor.

What is your question?

 

[solderdude]

All I asked is a test to measure how output ohms varies with frequency. on the preamp

Why would there be a variation impedance in frequency response within and well outside of the audible range ?
What would be the indicators/reasons for doing such a test ?
This is not a tube pre-amp where one could expect this kind of effects.

 

[Noske]

Why would there be a variation impedance in frequency response within and well outside of the audible range ?
What would be the indicators/reasons for doing such a test ?
This is not a tube pre-amp where one could expect this kind of effects.

I did not mention anything about audible range. Amir, however, did suggest that certain frequencies may be attenuated. I am sure that you have read that.

Tests can be done simply because they can be done. I need proof. I admire your concession that these effects may be real. Why stop at tubes. Plenty of solid state capacitor coupled preamps out there waiting to be tested.

 

[solderdude]

Then you should specify in which loads you want Amir to do the measurements.

1k ? 10k ? purely resistive ? SE and balanced ?

regarding your remark:

Amir, however, did suggest that certain frequencies may be attenuated.

Yes but this in passive mode. I already explained why that might occur:

In passive mode, the preamp naturally reflects the superb performance of the Audio Precision analyzer. Note that it will change the loading impedance so may impact performance of other devices connected to it. To avoid this, you want to use the active mode which still produces excellent performance.

Amir did measure both active and passive mode showing both had the same response (100k load) between 10Hz and 90kHz.

Of course, in passive mode the output impedance will be higher (vary depending on volume setting) which could limit the max. FR when longer or unusually high capacitance interlinks are used. That's where the 'transparent' buffer circuit comes in though which is low enough in output R to drive about any power amp input.
Every passive controller with a volume control will have this issue. Also LDR and transformer coupled ones.

 

[Noske]

So, I guess there is no point in Amir testing the ohms across frequency. That's OK. I have said what I have said.

 

[solderdude]

So, I guess there is no point in Amir testing the ohms across frequency.

Not in this case.

It would also require specific tests and there are no standards for what a cable or load should be so if you want something tested it would have to be clear what the loads should be and preferably also why (a specific cable with a specific length) and amp or some 'worst case' scenarios which then would say nothing about how it would behave in a single and common situation as that would not be near such a 'worst case' scenario.

 

[Noske]

Not in this case.

It would also require specific tests and there are no standards for what a cable or load should be so if you want something tested it would have to be clear what the loads should be and preferably also why (a specific cable with a specific length) and amp or some 'worst case' scenarios which then would say nothing about how it would behave in a single and common situation as that would not be near such a 'worst case' scenario.

I am getting the impression that you underestimate the abilities of Amir. He conducted tests on the Saga some years ago using a prior analyser.

Since when has output impedance been dependant on the cable attached? I don't get it. I encourage others here to explain the veracity of what has been conveyed.

 

[solderdude]

I am getting the impression that you underestimate the abilities of Amir.

Nope, I know what I can expect of Amir and what the capabilities of his various test equipment is.

I am sure Amir can do your requested measurements. You would have to define what conditions your requested measurements would have to be.
What level, what frequency span, passive, active, the load conditions (resistance, capacitance, inductance) balanced or SE and then you could test for those conditions.
So it is up to you to specify the test and if Amir sees value in that test and it can be done acc. to standards and he still has it he might be persuaded to add such test.

The question is what do you expect to see and the relevance of it (into resistive loads with a small capacitance).

He conducted tests on the Saga some years ago using a prior analyser.

I know. What is the relevance ?

Since when has output impedance been dependant on the cable attached?

It isn't and never said it was.

I merely explained in more detail what Amir meant with his remark and under which circumstances that could occur.

 

[SCG]

No POV on this part of the discussion, only offer this up as potential clarification. Below is the output impedance measurement Amir did on the Saga Tube preamp in Active mode back in 2019 with his comments below:

Looking all the way to the right, the minimum impedance is 183 ohm. At the lower end of 20 Hz, impedance climbs to 3,130 ohm. Using the typical 10:1 rule, your power amplifier input impedance should be at least 30k Ohm to avoid frequency dependent/EQ changes. Put more simply, if you use lower input impedance power amplifier after Saga in active mode, your low frequencies below 500 Hz or so will attenuate. This may accentuate the rest of the spectrum you may like, or not.

I think Noske is asking if it's still possible/valuable for Amir to test this on the solid state, updated tech version Kara in active mode?

 

[amirm]

I have the amp packed to ship to Schiit folks. I am reluctant to measure impedances with AP because it doesn't have a native capability for this. You have to install boatload of software (National Instrument) to get the UI you see above. If there is value, I can run such tests but here, we know that impedance varies with pot position so there is nothing new to be gained. Use low gain active mode to have load independence.

 

[VMRKarnati]

Thanks for the review. How much to look for Kara $699 in comparing with Saga+ $399 purely as preamp, apart from headphone part..... if we exclude balanced options also considering Saga+ gives the tube option as well and all passive + active proportions available at both, Can Kara definitely make us feel Overwhelmed in comparison or Saga+ is good enough? Please advise.

 

[Rottmannash]

What issue should be addressed ? The output resistance of 75 ohm (SE) or 600 ohm (Balanced) ? Why would that be an issue ?

The headphone out circuit design does not seem to work well when loaded with a low impedance load.
So think 50 ohm and lower, it seems to work fine for higher impedance loads like 300 ohm or higher.

As the pre-amp is the same circuit but without the higher power continuity output stage and will only have to drive a few k ohm as the 'heaviest load' it seems to be no problem for the same circuit.

Of course, in passive mode the output impedance will be higher (vary depending on volume setting) which could limit the max. FR when longer or unusually high capacitance interlinks are used. That's where the 'transparent' buffer circuit comes in though which is low enough in output R to drive about any power amp input.
Every passive controller with a volume control will have this issue. Also LDR and transformer coupled ones.

Would these same (impedance matching) issues affect the Freya+ using the passive and JFET buffer stages?

 

[solderdude]

It would similarly effect all passive volume controls.

It is not the pre-amp itself that is the issue but the output resistance in combination with the cable capacitance that forms a low pass in passive mode.
For this to reach audible levels you would need certain conditions which could occur with passive volume controls.
It will be worst at -6dB attenuation where the output resistance would be the highest.
A 10k pot, driven by a low output R source would have an output R of 2.5k ohm.
When this is connected with long cables (say 10m) to mono blocks close to the speakers or some really fancy (weird) high capacitance interlinks.
Let's say that a really poor cable capacitance would be 500pF/m (most are well below that) you would get a 5nF capacitance (10m).
This will give you a 3dB point at 13kHz which would be really audible.
Most cables, however, will be in the 100-150pF range (or even lower), so around 1nF for 10m = 60kHz -3dB so at 20kHz -0.5dB or so. This would only be so at -6dB and the cutoff point would shift upwards when shorter cables, lower capacitance cables or more or less attenuation was present.
It would be worse for longer lengths or passive attenuators with a higher resistance. I have seen 50k or 100k pots being used.

When such extreme cable lengths are being used one really should not use any passive attenuator (or tube pre with an unusual high output impedance).
In case of the Kara you should use the active outs (75ohm or balanced 600ohm) and this would be a non-issue.

The effect above is what Amir was talking about and what triggered Noske response.
The Kara is an entirely different circuit than the tube circuit of the Saga.

In case of the Saga (in active mode) something else is the problem. Here the problem is caused by a coupling capacitor. When you connect a low input resistance amplifier to the output the capacitance in the Saga will form a highpass filter (lows will roll-off) which is audible and not cable length issue with treble roll-off.
This is not the case with the Kara and does not have to be the case with a J-FET buffer stage (can have a higher capacitance coupling cap).
In the Saga we deal with high voltages so capacitors will be lower in value (otherwise they would be too big).

 

[Rottmannash]

Thanks for the response. I've been using the Passive mode on the Freya+ both to listen to 2 channel and as a pass through for HT. Perhaps I should use the JFET stage. I'm using XLR interconnects from the Freya+ to the Purifi amps.

 

[solderdude]

It would depend on the length of the cable and capacitance/ft (or meter).

 

[Rottmannash]

It would depend on the length of the cable and capacitance/ft (or meter).

less than 6 ft. Monoprice cables.

 

[HoweSound]

I'm still waiting to hear what measured improvements Kara brings to the table over it's predecessor Freya S. It failed to cure my Freya complaints that I consider easy fixes: a front mounted on/off switch, a trigger function and a remote on/off.

 

[solderdude]

less than 6 ft. Monoprice cables.

Probably between 150 and 500pF max.
So with max. 2.5k output R the bandwidth is higher than 100kHz.... worst case.

 

[Noske]

I have the amp packed to ship to Schiit folks. I am reluctant to measure impedances with AP because it doesn't have a native capability for this. You have to install boatload of software (National Instrument) to get the UI you see above. If there is value, I can run such tests but here, we know that impedance varies with pot position so there is nothing new to be gained. Use low gain active mode to have load independence.

Thankyou. I acknowledge that given the circumstances, the measurements cannot be completed.

Whether they were would have been interesting or not remains an unknown. I am very pleased for the necessary results provided, of course, thankyou, but sufficient?