Help with output/input impedance matching and high frequency loss

[Tilock]

Thanks in advance for any help or advice you can offer!

My speakers, Merlin VSM-MXr come with a custom hardware bass EQ that bumps them up +4db at 35hz and rolls of the bass more quickly. It goes between my yamaha CD-2100(high end SACD player/DAC) and my balanced 300B set tube pre amp. I'm integrating a sub for the first time via the second set of pre-out from the line stage and removed the EQ so that I don't boost and roll off the sub but when I ran frequency sweeps(REW, calibarted mic) before and after there was a noticeable drop out in the high freq starting at 2khz and peaking at -4dB at 10khz when I was running direct from CD player to pre-amp.

The EQ has a stable output impedance of 100ohms and I measured the output impedance of my CD player to be 530ohms(seems quite high but I verified it multiple times). I contacted Yamaha support and they verified that this seemed correct although it's not a published spec. I tried the XLR(measured at 1050ohms) out on the CD player but it was even worse with 15dB down at 10khz.

So it appears that my pre amp cannot handle the output impedance of the CD player causing high frequency loss. I measured the output voltage before and after the EQ and it's the same but if I remove the EQ the output voltage at with a 10khz sine signal drops from the preamp. The pre-amp is supposed to have an input impedance of 100,000 ohms. I verified using another source with lower impedance and without the EQ that there's no high frequency drop out. I also made a test tone cd and verified this happens both through the USB DAC and CD only output. This and the fact that the problem got worse with the higher impedance XLR suggest strongly this is an impedance based problem to me. To get the voltages and impedance numbers I used REW to generate a sine wave at various frequencies and used the USB input to my SACD player's DAC. Attached aligators to the RCA/XLR cables and used a multimeter(using a known resistor to load the output so I could calculate the impedance).

So basically I'm wondering is 530ohms very high output imp for a CD player and should a pre-amp with 100,000ohms input impedance struggle with that at all? I don't have the experience to know if 530ohms is abnormally high and should be looking for a problem inside the CD Player or concentrating on the pre-amp. What is generally the desired ratio for output/input impedance?

Is there a relatively easy way to measure the input impedance on the pre-amp to make sure that it is actually 100k Ohm as advertised? I'm suspicious that it may not be as I also tested the CD player with a Rotel integrated(with pre outs) which only has 47K ohm input impedance and there was no drop out.

I may be missing something obvious due to my lack of experience so I'm happy to have the assumptions I've made corrected.

Thanks for your time and consideration!

 

[rayma]

The 500 ohms is not high for a source component output impedance, and is fine. Any load impedance over 10k is fine.
Have you checked each component individually so an equipment fault can be ruled out?
Perhaps there is something wrong with your measuring setup.

 

[Mooly]

So basically I'm wondering is 530ohms very high output imp for a CD player and should a pre-amp with 100,000ohms input impedance struggle with that at all? I

It sounds a bit of a complicated setup but I can say for sure that 530 ohms is 'low' in the scheme of things and should drive any preamp with ease. 100k input impedance is a typical figure neither very high nor low.

530 ohms driving 100k reduces the level by a tiny tiny fraction of a db

To roll the HF off from that CD player would imply (and this is highly unlikely in normal circumstances) that it is driving a large capacitive load. That capacitance would need to be massive in context with what we normally see regarding inputs. 20nF would give an audible roll off for example but typical values normally seen would be around 20 times lower than that (2nF and below).

So whatever is going on, the output impedance of the player is not to blame.

Is there a relatively easy way to measure the input impedance on the pre-amp to make sure that it is actually 100k Ohm as advertised?

What equipment have you?

A scope check (or accurate meter check of signal level ) looking at how much the preamp pulls a test signal from the CD player down when the preamp is connected would allow the impedance to be calculated.

 

[Mooly]

The output impedance is actually the value of these two resistors added together and is the same for RCA and XLR feeds. So 540 ohms according to these.

(it all looks very Marantzy to me, the circuit and the way its all drawn)

 

[Tilock]

The 500 ohms is not high for a source component output impedance, and is fine. Any load impedance over 10k is fine.
Have you checked each component individually so an equipment fault can be ruled out?
Perhaps there is something wrong with your measuring setup.

Yes, I've verified it a few different ways. The frequency fall off is definitely there and goes away with either the EQ(100ohm) or a lower impedance source(using the line out from the computer instead of USB out through the CD player). I can also see the voltage from the pre-amp drop off when fed a 10khz sine wave from the CD player but not from the lower impedance sources. Thanks for your response.

 

[rayma]

Are you using a long, high capacitance cable to connect the components?
It would need to have as much as 20nF shunt capacitance to behave as you describe.
Are you using an uncompensated probe for the oscilloscope?

 

[Tilock]

Moody, no scope but I do have a fairly nice meter which is what I used to calculate the output impedance from the CD player, line out and verified that that EQ was actually 100ohms.

I have a standalone USB DAC on the way which has an output impedance of 200ohms so I'll be able to test that to see if there's any drop off with that device.

Thanks for the circuit diagram! I appreciate it.

Does the fact that the line out (1/8th headphone jack to RCA) was able to drive the pre-amp fine with no high frequency drop out change the possibility of there being a large capacitance load on input of the pre-amp?

Could you explain in a little more detail how I can test the input impedance of the pre-amp with a multimeter and test signal? I can measure how much it changes the from output of the CD player and then the output of the pre amp when I output a 10khz sine wave is this what you're suggesting?

 

[Tilock]

@rayma No, a 3 foot Cardas Golden presence interconnect and I've also tried multiple cables with the same result. I don't have an oscilloscope. I've been using a multimeter to calculate the voltage/impedance and a calibrated microphone to get the frequency response.

I've also measured each channel separately and they both experience the drop out.

Also the output voltage of the test signal does not change the frequency response(as in if I set the test tone at 0 dB or -12dB and change the volume on the pre-amp to match levels).

 

[rayma]

Does the signal source itself not show any high frequency drop off?
Many (most) meters lack response above 1kHz, so test that first.

Don't use a mic to attempt to measure an electrical response, since that will be a big source of error and also hard to repeat.
Just use the meter, but verify that the source remains constant with frequency.

 

[Tilock]

@rayma No. The source shoes no high frequency drop off when it's run through the hardware EQ first and the voltage measure on input and output is near identical before and after the EQ. The EQ does not alter the high frequency. In this case it's just changing the 530ohm output impedance of the CD Player to 100ohm.

I'm only measuring the voltage and impedance with the meter. The frequency measurement are taken through my speakers with calibrated microphone. So when I test I send a tone through my laptop which is connected to my cd player via the USB DAC inside the player. The only change I made when noticing the drop off was removing the hardware EQ from the chain. So CD PLAYER - EQ - PRE-AMP - MONO BLOCK - SPEAKER or CD PLAYER - PRE-AMP - MONOBLOCK - SPEAKER.

If I don't move the microphone repeated measurements are accurate to less than 0.5dB.

When I changed the pre-amp the problem goes away, when I change the source to the line out on the laptop directly(much lower impedance) the problem goes away.

 

[rayma]

So the loss only happens with the tube preamp in the chain?

 

[Tilock]

@rayma

Yes, but only from the higher impedance source. If I use a lower impedance source the problem goes away(EQ or same laptop but skip the CD player by using the line out and laptops internal DAC which has an output impedance of under 100 ohms).

Just took some measurements. 10khz test tone 1v output from laptop into USB DAC in CD player : 0.450v output from CD player into EQ - 0.442V output into pre-amp output - 0.331v into power amps.

If I remove the the EQ and go direct from CD PLAYER to Pre amp(fixed volume). CD OUTPUT 0.450v - Pre-amp Output - 0.215V.

So using the high impedance output of the CD player direct to the pre-amp results in a voltage loss of 0.116v at 10Khz.

Thank you for taking the time to work through this with me.

 

[rayma]

The thing is, source impedance of as much as 1k is common and not at all problematic, except in very unusual circumstances.
For -3dB at 20kHz and 1k source, the capacitance would be ~10nF. So there's something unusual about your preamp, apparently.

Unless you have a schematic, or can reverse engineer it, we can only assume that this preamp has abnormally high input capacitance, which places an undue burden on the source.

 

[Tilock]

@rayma I agree. I think there is something wrong with the pre-amp and I'll have to look deeper into exactly what it is. It's all point to point wired and all the components are identifiable. I'm not knowledgeable enough to just look at it and see the issue but I should be able to at least map out the circuit.

This is going to be a weird question but could having a TCV set to max and fixed on the input side and a ALPS Blue 50k ohm pot controlling volume on the output side cause this issue?

 

[rayma]

Can you draw a diagram? This is in your preamp?
The transformer control will certainly have different source loading than a pot, more like a parallel LC.

Also some sources will have trouble driving a TVC load, with such a small DC resistance.
If those controls are in the chain, remove them and try again.

 

[Tilock]

@rayma Yes, it's my pre-amp but I did not build it. The dual volume design was a fix from the manufacturer due to problems with the initial design using the TCV only. I can get pictures of the circuit tomorrow. Thanks.

 

[rayma]

Then at the input of the tube preamp there is a TVC, then the 300B, then a standard pot.
This is rather unusual, I think.

 

[Tilock]

@rayma Oh, I agree. Yes. TCV connected to the input transformers now fixed at full output and a alps blue pot on the output for volume control. I was not informed what the fix was before the amplifier was returned to me.

 

[rayma]

So actually at the input of the tube preamp there is a TVC, then an input transformer, then the 300B, then a standard pot?
We really do need the schematic.

 

[Tilock]

Well, it's my understanding that the TCV set at max and locked in place is tapped directly to the input transformers(it's a fully balanced design so there's one for each channel) in front of the 300B and then post 300B the dual ALPS control the gain before the output.