TRL Dude or Joule 150 MKII for Major Pre Upgrade

Ralph, the present (original) value is 100 ohms. Do you still feel that is not the problem?

Best regards,
-- Al

Tim (Mitch2), Ralph's comment that pin 3 should be grounded when adapting to an XLR input using a simple RCA-to-XLR adapter is of course correct. I would add that the 23.5K nominal input impedance of your Jensen PI-RX transformers is specified based on a 20K load being connected across its secondary. See the datasheet. With a 100K load I would expect its input impedance to be, as a rough approximation, more like 100K!

A separate question would be how well the transformer will perform when loaded with 100K and driven by the Dude's output impedance (I don't know what that is). I note that the datasheet indicates a maximum allowable load of infinity ohms, but a maximum allowable source impedance of 2K. A conversation with Jensen would seem to be in order on that question.

An ideal transformer (which of course is something that does not exist) having a turns ratio of one-to-one will present an input impedance identical to the load impedance that is connected across its secondary (output) winding.

Best regards,
-- Al

Tim, sounds like the single-ended preamp has a ground loop problem when working into the balanced amp via an adapter, the loop being broken, as might be expected, when the transformer is used.

An adapter-based approach that I suspect would provide better results, although probably not as good as the transformer-based approach, would be to connect RCA male to XLR male adapters at the preamp outputs, and run XLR cables from there to the amp. That would essentially be equivalent to the arrangement shown in Figure 2.1 of this Jensen paper on interfacing balanced and unbalanced equipment.

With respect to using the transformer, I wouldn't necessarily conclude that you have to stay below 2K output impedance for the preamp, given that your amp has an input impedance of 100K. It seems conceivable to me that the 2K figure in the datasheet is based on the assumption of a 20K load, just as the input impedance spec is based on that assumption.

However, when deciding on whether to purchase this single-ended Dude, I called them again this past week and was told the input impedance would be a maximum of 25K ohms, as shown on the data sheet.

My guess is that the person you were speaking with was not Mr. Whitlock :-)

Maybe another phone call is in order.

Yes, that would seem clear.

On the other hand, you could buy one of Ralph's preamps, and the issues would all be moot :-)

(I have no affiliation).

Best regards,
-- Al

Grannyring, your issue doesn't seem to me to be a ground loop issue, as it occurs even when the preamp is unplugged from the AC outlet (and with no source components connected). Therefore I'm not at all certain that a Jensen transformer would solve the problem, which I'm frankly out of ideas on.

For anyone else who may be able to contribute ideas on Grannyring's issue, it was discussed here. Note that multiple sets of interconnect cables were tried, as well as both the Dude and Lightspeed preamps. Yet the problem goes away when the inputs to the amp are disconnected, and the problem did not occur until the previous speakers, which were electrostatics, were replaced with passive speakers.

Best regards,
-- Al

Gentlemen, once again let me emphasize that Grannyring's issue arises even when the preamp's power cord is not plugged into an AC outlet, and when no components are connected other than the preamp, power amp, and speakers. Therefore I don't see how the problem could be due to a ground loop, because there is no loop in that situation.

The facts that the problem is not present when nothing is connected to the inputs of the amp, but is present when the preamp and interconnects are connected to the amp, would seem to suggest an issue involving emi/rfi pickup. However, as Grannyring indicated in his thread that I linked to earlier, in addition to its normal RCA and XLR inputs his amp provides separate RCA and XLR inputs that are processed through high pass filter circuitry. And when the preamp is connected to the normal RCA inputs the problem exists even if any of the other three inputs are selected on the amp. Which leaves me totally baffled, unless perhaps something is defective in the amp's input select circuitry. But that, in turn, would not seem to explain why the problem wasn't present with the previous speakers.

Vett93, regarding your otherwise excellent posts it should be pointed out that with many designs isolating signal ground from chassis and AC safety ground may involve considerably more than minor modification. For starters, the ground sleeves of RCA connectors may not be isolated from chassis. Also, it would seem possible that in many cases changing the component's internal grounding scheme might result in unpredictable sonic side-effects.

Regards,
-- Al

This was extremely bizarre to me...

To me as well.

Have the preamps been located close to a possible source of emi/rfi, including perhaps the amplifier itself? If that seems like a possibility, try moving the preamp to different physical locations.

Beyond that, I'm out of ideas at this point.

Best regards,
-- Al

Sounds like Vett93 gave you some excellent suggestions. One additional experiment I would suggest is that besides shorting the two pins on the XLR connectors, you also short the center pin and ground sleeve of the two RCA connectors for the high passed inputs of the amp. If you don't have shorting plugs, you could again use IC's with something conductive inserted at their other end to provide the short.

I would add to your summary, btw, as you mentioned in the other thread where this was discussed, that the noise is present if the preamp is connected to the regular inputs of the amp and the high passed inputs are selected, but disappears if the preamp is connected to the high passed inputs and the regular inputs are selected (while being present if the preamp is connected to the high passed inputs and the high passed inputs are selected).

And yes, the caution about not grounding the negative output terminals of the amp is not applicable if all that is connected is a passive speaker. It shouldn't be applicable with an electrostatic, either.

Best regards,
-- Al

I wouldn't expect that shorting the preamp inputs would make any difference, since the problem occurs even when the preamp is turned off and unplugged from the AC.

Sure is a baffling problem :-)

Best regards,
-- Al

Grannyring, what kind of speaker cables are you using, and how long are they? If by any chance they are one of the kinds that has extremely high capacitance, such as Goertz or Ocos, and you are using them without a Zobel network, I'm wondering if that may be resulting in the amp being on the edge of having some sort of low level oscillation, that could somehow be put over the edge by conditions at the inputs.

Also, did the positioning of the speaker cables change significantly when you changed speakers? Perhaps they are picking up rfi/emi which is entering the feedback loop of the amp, if it has one.

Just some shots in the dark.

Best regards,
-- Al

I agree completely. 100 ohms sounds way too low. And as the link you provided indicates, too low a value for the grid stopper resistor of the input stage could very conceivably result in rfi problems.

According to some quick calculations, 1K strikes me as a reasonable value. I believe that a value up to around 4K or 5K would not adversely affect anything, the overall bandwidth of the amp in particular.

Kudos to Johnsonwu.

Best regards,
-- Al

No, don't solder a resistor between the RCA and pin 1 of the tube socket. That would put the 1K in parallel with the 100 ohms, resulting in an overall resistance of about 91 ohms.

Hopefully you can unsolder the wire that is presently connected to the RCA jack, and put the resistor in series (i.e., connected between that wire and the RCA jack). That would result in an overall resistance of 1100 ohms.

It might be a surer bet, though, to use a resistor having a higher value, such as the 3K or 4K values I mentioned. The main constraint on how high a value should be used is, I believe, that if it gets too high the bandwidth of the low pass filter formed by that resistance and the input capacitance seen looking into the grid of the tube, based on the Miller Effect, will reduce the overall bandwidth of the amplifier (which is spec'd at 150 kHz). Calculating from the 6SN7 data in my tube manual, I believe that 4K would still be low enough to not have a significant effect on that bandwidth.

Best regards,
-- Al

Success, at long last. Outstanding! Although I guess we'll never really be able to be sure of exactly what was causing the problem, given the bizarre results of some of the experiments.

The key thing with the XLR cable is that its capacitance has to be kept very low. I suspect that you are using the PI-2RX transformer, for which the datasheet indicates a maximum acceptable load capacitance of 100 pf. That would be for the sum of the cable capacitance and the input capacitance of the amp (which is unspecified). I suspect that the main consequence of driving the transformer from an output impedance that is significantly higher than the 2K recommendation will be that the load capacitance becomes even more critical, excessive capacitance potentially affecting the highs particularly.

For most purposes the capacitance of Mogami's Gold Studio or other Mogami 2534-based cable (if that is what you are using for the time being) is reasonably low, but it may not be low enough in these circumstances. I don't know what kind of cables you are awaiting, but if they do not have very low capacitance you might want to consider say a 1 foot length of inexpensive Blue Jeans/Belden 1800F, which would be only 13 pf plus the capacitance of the connectors.

Best regards,
-- Al

Ralph, thanks for the info about 100 ohms. I would not have suspected that such a low value would generally be effective.

Regarding the effects of the transformer on the source impedance seen by the amplifier, though, I don't think that is the explanation. For two reasons:

1)It was reported that the problem was still present when the amp was driven by several different low impedance sources.

2)The Jensen Isomax "input transformers," such as the one whose data sheet I linked to in my previous post, do not provide a low output impedance. That presumably being a major reason that they are only recommended for use with short low capacitance cables on their output side. As you can see in the data sheet I had linked to, that particular model has an output impedance spec of 4.65K (typical) when it is driven from a balanced impedance of 300 ohms per leg. Also, the DC resistance of its secondary winding is spec'd at 1.9K (typical). When driven by a source having high output impedance, such as the Dude, the 4.65K figure would presumably be even higher.

So as I see it the mystery remains unsolved.

Best regards,
-- Al