A Question of Impedance

Housteau · « **on:** 23 Jan 2009, 09:58 pm »

Most tube preamps like to see an input impedance of at least 20K on the amps that they drive. There are some amps that when being driven together in parallel can drop below that number. Is there a simple technical solution around this? I have seen line level isolation x-formers recommended for different situations between components. Could that be done here between a preamp and an amp with a low input impedance? What sort of impedance would the preamp see from such a x-former?

richidoo · « **Reply #1 on:** 23 Jan 2009, 11:44 pm »

Many tube preamps can handle 10kOhm load without a problem, so try it out before considering a fix. Even a good simple design like Minimax drives 10kOhm to ~19V (iirc) before clipping on a single 12AU7 buffer tube.

Housteau · « **Reply #2 on:** 24 Jan 2009, 05:58 pm »

My Audio Research and Monarchy Audio both state 20K minimum, but as you say they may tolerate a lower number. I am a planner and away from home at this time. So, for now I am assuming that there will be issues and I am looking for alternative solutions.

JoshK · « **Reply #3 on:** 24 Jan 2009, 06:49 pm »

The step down transformer would work. Impedances are the square of the turns ratio, but the voltage is proportional to the turns ratio, so let's say you have a 2:1 step down. You will half the gain of your preamp but the output impedance of your preamp is quartered.

The rule of thumb is that the input impedance (Zin) of the amp should be atleast 10 times the output impedance (Zout) of your preamp. So if you cut the Zout by 4 then you could push an amp with a Zin of 5K. Or in your case the combined load of all the amps in parrallel.

All this ignores the driving current of your preamp, so far. IMO, there is more to driveability than ratio of Zs. Many SS amps have a lot of input capacitance which roll off the bandwidth of your system if your preamp doesn't have enough drive current. Plate amps are worse. The good news is that with output transformers, the gain that you halved in this example is traded for twice the output current.

How much extra gain do you have? Where on the dial (from where to where) do you listen to and how about the max? If you were to step down the gain 4:1, then Zout gets divided by 16 and you have 4 times the output current, gain gets quartered though.

Typically tube preamps have too much gain when used with SS amps. How much gain does your preamp and amp have.

Housteau · « **Reply #4 on:** 24 Jan 2009, 09:25 pm »

That is some very good information. My preamps are tube and my normal listening would place them in roughly the 09:00 position. I don't currently have a problem, but that is because I cheated in how I have my system hooked up. I basically have a passive triamped system with VTL monos on top, an older Electron Kinetics Eagle 2a on bass down to 57Hz, and mono Dayton amps on the sub bass. The VTLs and Eagle are in direct parallel as they have normal input impedances. However those Daytons (which are great for low bass) have an input imp. of 12K. Those amps have a separate gain contol and so I used a bridging circuit to place them in the system. I run them off the output from my VTLs using a voltage divider and dropping the signal back down to line level. Since it is just the low bass this works just fine.

However, now I am planning to go active instead of passive. Since the Dayton has its own gain control I can run it in parallel with the active x-over that will feed the high and mid bass sections. That x-over has a low input impedance itself of around 20K. The other option is to run the Dayton in parallel with the Eagle 2a from low output from the x-over.

Why would I not just do another bridging circuit? Because of the active x-over I would need to run it off of the outputs from the Eagle 2a since the VTLs are only receiving the higher frequencies. The problem is that this old Eagle has a hum, ok by itself, but it gets amplified then input to the Dayton. Then it becomes noticible.

JoshK · « **Reply #5 on:** 24 Jan 2009, 09:45 pm »

It sounds like you have a good handle on the situation.

It also sounds like you have plenty of gain to spare. Reducing that gain will actually improve the S/N ratio of your system and should make it sound better as a result. So the transformer idea is not a bad one.

Since you likely have caps in the output of your tube preamp, I'd recommend a parafeed output transformer, like B7s from Magnequest (about $60/ea). I'd wire them up into a pair of IC's. Put it close to the preamp end. Then you can actually experiment with and without. If you really like the result then you can permanently wire it into the preamp.

A note about picking the right B7, since you have choices. The B7s are quoted in impedances. You most likely want the 5K:500. 5,000:500 is 10:1 impedance ratio, but turns ratio is the square root of this, or ~3.2:1, so voltage gets stepped down 3.2 and current gets multipled by 3.2. The Zout is 1/10th the original though. The 10K:500 is roughly 4.5:1 voltage ratio, 20:1 impedance ratio, but I recommend the 5K:500.

Said another way, if the load you are driving is 12K, if you are using a 5K:500 impedance ratio, then the preamp now sees a 120K load instead of the 12K load. Transformers are cool huh?