Ref 5SE in Stereophile -- higher impedence amps?

[pdubya]

In the Measurements section of November's Stereophile review of the Ref 5SE (no link yet) JA suggests that the preamp should be paired with "higher impedence amplifiers". What is the general demarcation between high and low impedence in amps? What would the results of a mismatch sound like?

I currently listen to Salon2s by way of an ARC LS27 driving Classe CT-M600s, which have an input impedence of 50k?. ARC's own Reference amps' input impedences range from 200k? to 300k?. I'm considering upgrading to the 5SE. I typically use the LS-27's Low Gain setting (+12 dB), which is equivalent to the 5SE's fixed gain. Even though I keep a Post-It with volume settings for each disc I find I'm having to make adjustments in the course of listening way more than one would expect (and it's not just always turning it up due to encroaching deafness ...). Would this be a result of mismatching?

I would appreciate any illumination before upgrading to more of the same.

 

[DaveyF]

I just finished reading that review. What amazed me was the disclosure from JA, wherein he stated that the unit was damaged with a major cap having broken off! He was unable to correctly solder the cap onto the board, so he kluged the cap with clips and took the measurements....WOW. Surely, it would have been prudent for ARC to send him a new unit... after all they sent him the pieces for the so-called fix! With that said, I would suggest that you try and listen to the 5SE in your system and NOT rely on that review or anyone else's opinion of the preamp's sound..just IMHO.

 

[DonH50]

SS amps typically have 10k ohm inputs. Tube amps typically 100k ohms. More or less. Tube preamp outputs are often very high impedance, like 1k or more, compared to 100 ohms or less for SS preamps.

 

[microstrip]

Audio Research usually recommends using amplifiers with XLR input impedance greater than 20 kohms with their tube preamplfiers, although some people report good results using input impedance lower than this value. We also have to consider that the specific advised value for each particular case will depend on system - if you have a high sensitivity speaker with an amplifier with high gain, the maximum amplitude delivered by the preamplifier will be much lower, allowing a lower input impedance.

 

[MylesBAstor]

SS amps typically have 10k ohm inputs. Tube amps typically 100k ohms. More or less. Tube preamp outputs are often very high impedance, like 1k or more, compared to 100 ohms or less for SS preamps.

Not all tube preamps Don. For instance the GAT's output impedance is 100 ohms, in part since they forsaked a traditional cathode follower.

SS amps will vary but I haven't run across in my travels many SS amps with 10 kohm input impedances. Most are like 20 Kohms.

 

[DonH50]

I took a quick look. I am sure there is a wide range. My old SP3a1 was rated for 20k, I think, but was a lot happier with 100k.

 

[mep]

Not all tube preamps Don. For instance the GAT's output impedance is 100 ohms, in part since they forsaked a traditional cathode follower.

SS amps will vary but I haven't run across in my travels many SS amps with 10 kohm input impedances. Most are like 20 Kohms.

Cathode followers are normally used to lower the impedance. If CJ didn't use a cathode follower, did they use some type of SS buffer to reduce the impedance?

 

[LL21]

Cathode followers are normally used to lower the impedance. If CJ didn't use a cathode follower, did they use some type of SS buffer to reduce the impedance?

I think that is the case, yes. From their website:

"The circuitry of the GAT Preamplifier is elegant in its simplicity. The audio circuit consists of a single amplifier stage incorporating one composite triode comprised of paralleled sections of a high-transconductance miniature dual triode vacuum tube. This amplifier stage is coupled to the preamplifier outputs through a high-current MOSFET buffer, which achieves a very low output impedance, making the GAT extremely flexible in the choice of interconnect cables and amplifiers. DC voltage is supplied to the circuit by our latest discrete voltage regulator that isolates the audio circuit from the power line by maintaining negligible impedance across the audio frequency band. Infrasonic noise is minimized by operating the tube filaments (heaters) on a dc voltage supplied by another discrete regulated power supply."

 

[MylesBAstor]

I took a quick look. I am sure there is a wide range. My old SP3a1 was rated for 20k, I think, but was a lot happier with 100k.

A lot of older tube preamplifier designs did have very high output impedances. But designers have been working at dropping that number over the years for reasons we all know. I think the ART preamp with a cathode follower was 500 ohms.

 

[DonH50]

I cannot recall if my ARC had a cathode follower or not. I did build one for a Dynaco but IIRC the output impedance of the cathode follower was still pretty high, maybe in the 1k ohm range. Since I was driving a tube amp it did not matter. One of the custom preamps I designed had paralleled cathode followers to hit the 100 - 200 ohm range.

I used to pad a tube preamp's outputs fairly routinely for friends with SS amps. Most tube preamps had enough gain and gobs of voltage headroom, so adding a resistor of 10k - 20k in series with the output didn't do anything really bad, cost about 3 - 6 dB (turn up the volume) and generally measured and sounded a lot better to me.

IIRC, I measured 50+ Vpp out of my SP3a1, talk about some output swing!

 

[microstrip]

I cannot recall if my ARC had a cathode follower or not. (...)

Older ARC tube preamplfiiers since the SP3 had the same topology - a gain stage using two triodes and a cathode follower. However they had a 475 or 511 ohm resistor in series with the output capacitor.

 

[MylesBAstor]

I cannot recall if my ARC had a cathode follower or not. I did build one for a Dynaco but IIRC the output impedance of the cathode follower was still pretty high, maybe in the 1k ohm range. Since I was driving a tube amp it did not matter. One of the custom preamps I designed had paralleled cathode followers to hit the 100 - 200 ohm range.

I used to pad a tube preamp's outputs fairly routinely for friends with SS amps. Most tube preamps had enough gain and gobs of voltage headroom, so adding a resistor of 10k - 20k in series with the output didn't do anything really bad, cost about 3 - 6 dB (turn up the volume) and generally measured and sounded a lot better to me.

IIRC, I measured 50+ Vpp out of my SP3a1, talk about some output swing!

As I understand it, the output impedance affects among other things, the preamplifier's ability to drive long runs of cable.

 

[microstrip]

As I understand it, the output impedance affects among other things, the preamplifier's ability to drive long runs of cable.

It will depend on the capacitance of the cable. We can have very low capacitance cables, with as little as 20 picofarads per meter and I remember reading about cables with ten times higher higher capacitance per meter. It is why we sometimes find very contradictory opinions on this subject about a specific preamplfier.

 

[Steve Williams]

It will depend on the capacitance of the cable. We can have very low capacitance cables, with as little as 20 picofarads per meter and I remember reading about cables with ten times higher higher capacitance per meter. It is why we sometimes find very contradictory opinions on this subject about a specific preamplfier.

is it low capacitance or rather high inductance cables

 

[MylesBAstor]

It will depend on the capacitance of the cable. We can have very low capacitance cables, with as little as 20 picofarads per meter and I remember reading about cables with ten times higher higher capacitance per meter. It is why we sometimes find very contradictory opinions on this subject about a specific preamplfier.

Exactly. For instance, IIRC, the original vdh Carbon cables were very high capacitance and caused my friends Classic 150 to blow up. So were those old, infamous Cobra cables.

 

[DonH50]

A cable is a distributed RLC network (one way of looking at it). For interconnects, R is usually too small to matter since the load is high-impedance, and cable inductance is usually low enough that C dominates. I have always looked for low-C cables for interconnects, especially with tubes. Typical interconnects use 75-ohm video cable that runs around 20 - 30 pF/foot. Cheap cables may run a little higher, and the cables I used for my tube gear were around 10 pF/foot (I forget exactly the coax, think it was 93 ohm or something not terribly usual these days). Myles is correct about higher C affecting the bandwidth for long runs. Realistically none of the parasitic RLC matters at audio frequencies for normal runs (maybe 10' - 30' or less, there's a thread on interconnect frequency response in the WBF technical area). It is more a factor in live sound where you may have a mic snake to a mixer 50' to 100' from the stage, one reason balanced XLR rnus are used and lower-Z (600 ohm) impedances.

I was at the CES where the Cobra cable guy went around rooms trying to get people to try them, and saw the Infinity switching amp blow up. Not a pretty thing (although there was no smoke, the amp just went "ziiip" and dead). Amps rise in impedance with rising frequency, and some do not handle capacitive loads well (become unstable). Cobra cables and similar woven designs reduced inductance but the trade is very high capacitance. In my mind a flat ribbon would provide the best trade of low L and low C with low series R as well.

I agree with microstrip that cable differences we hear are the effect of the cables interacting with the components, and with Frank van Alstine that any such interaction we hear is likely the sign of a poor design. I am a little less harsh on the latter knowing the design trades that might influence higher output impedance.

I have not thought much about speaker cable impedances save the R (which matters more due to the low impedances involved on the speaker side of the amp), though am pretty sure one of my techie articles here on WBF does delve into that.

HTH - Don