2. amps-preamps
  3. 1349886755

1.2k Output Impedance to 10k Input Impedance

I have a tube preamp with only single ended outs that does not have any output buffers and therefore a fairly high output impedance of 1.2k. My SS power amp has input impedances of 10k on RCA or 20k on XLR.

I have been advised by the designer of the preamp that anything lower than 50k or so will start loading down the line stage, so it would seem an imperfect pairing. My assumption is that at the very least I’d want to use the XLR inputs on the amp, as they present a 20k load to the preamp instead of the 10k vis the RCA inputs.

I have an RCA-XLR interconnect, but I’ve been told that simply using this will not result in a 20k load on the preamp and that the only way to accomplish this would be to use a passive transformer in line, such as the Jensen ISO-Max DM2-2RX.

For those that have experimented with these devices, do they result in any sonic degradation?

I'm wondering which might be the lesser evil – introducing an additional circuit and set of connections into the signal path via the DM2-2RX transformer, or running with a less than ideal impedance match between preamp and power amp. The answer is, of course, to try both and see, I'm just wondering if anyone has gone down this path before and what they found (heard).

Any inputs would be greatly appreciated.
srosenberg

Showing 2 responses by almarg

almarg's avatar
almarg
9,670 posts
 
I spoke to Jenson and was advised that the most appropriate unit for my needs would be the PC-2. it is a 4:1 ratio and therefore would satisfy the needs of both the preamp and poweramp - the preamp seeing a load of 34k ohms and the power amp 225 ohms.... The PI model that you mention would result in a 23.5k ohm load on the preamp and 4.65k ohms on the amp, which if i'm understanding correctly, is less than ideal (ratio of 5:1)
It's not quite that simple, although the conclusion is essentially correct.

The input and output impedance numbers shown in the Jensen datasheets are based on specific source and load impedances ("test circuit 1" shown at the bottom of page 2 of the datasheets), and will vary considerably depending on the output impedance of the preamp and the input impedance of the amp. In this case the balanced input impedance of your amp happens to correspond to the test conditions shown in the datasheet for the PI model, so the 23.5K figure is applicable. But the output impedance of your preamp is much higher than for the test condition shown in the datasheet, so the source impedance seen by the power amp will be considerably higher than 4.65K.

It is important to note, though, that the key concern is not the relation between those two numbers, but the relation between the load impedance seen by the preamp (23.5K in this case), and the actual output impedance of the preamp, at the frequency for which it is highest. See the last paragraph below for further theoretical explanation.

The numbers for the PC-2 (you would want the PC-2RX specifically, assuming that you want to use the XLR inputs of the amp) are based on source and load impedances that differ even more widely from those of your components, and also differ widely from the test conditions of the PI model. That said, though, the PC-2 does appear to be much better suited to your purposes, as does Steve McCormack's Interocitor.

Keep in mind, however, that a volume reduction goes hand-in-hand with the impedance transformation you are trying to achieve. That is specified as 12 db for Steve's unit, and typically 13.6 db for the PC-2. If without the transformer you would be using the volume control in the upper third or so of its range (roughly speaking) that would be a concern. On the other hand, if (as is common these days) you have too much overall gain in your system, and use the volume control in the lower part of its range, that would be beneficial.

The theory involved in all of this is that for an ideal transformer (no practical transformer is ideal, "ideal" meaning among many other things that the DC resistance of the windings is insignificant) the ratio of output voltage to input voltage corresponds to the ratio of the number of turns in the secondary (output) winding to the number of turns in the primary (input winding), while the impedance transformation corresponds to the square of that ratio. It appears that both the PC-2 and the Interocitor have turns ratios of roughly 4 to 1. So the power amp would see an input voltage roughly equal to 1/4 of what the preamp is putting out, which corresponds to a 12 db reduction. The preamp would see a load impedance that is roughly 16 times as great as the input impedance of the power amp. The power amp would see a source impedance that is roughly 1/16th of the output impedance of the preamp. I use the term "roughly" to take into account DC resistances of the windings, various other causes of "insertion loss," variation of the actual turns ratios from the 4:1 nominal value, etc.

Best regards,
-- Al
almarg's avatar
almarg
9,670 posts
 
Hi Scott,

IMO, the impedance matching guideline is properly stated as follows (I'm quoting from myself in another thread):
Ideally the input impedance of the power amp should be at least ten times greater than the output impedance of the preamp, at the frequency within the audible range for which the output impedance of the preamp is highest. That frequency will usually be 20 Hz, as a result of the impedance rise caused at low frequencies by the output coupling capacitor most tube preamps use.

A factor somewhat less than 10x may or may not be acceptable, depending mainly on how the preamp's output impedance varies as a function of frequency, and also on the deep bass extension of the speakers.
The fact that the designer suggests 50K minimum tells me that the preamp's output impedance probably varies considerably as a function of frequency. And that it is much higher at some frequencies, probably including 20 Hz, than the 1.2K nominal figure (which I suspect is at 1 kHz).

The 20K balanced input impedance of the amp probably represents the sum of 10K input impedances on each of the two legs (i.e., on each of the two signal inputs that comprise the balanced signal pair). So using an RCA-to-XLR cable will not help, as you were advised.

Concerning the Jensen transformers, member Mitch2 has an excellent system and has spoken very enthusiastically in past threads about the results he has obtained with them. Atmasphere, on the other hand, has commented that in his experience they do have at least some slight artifacts.

Best regards,
-- Al
More to discover