Bridged-T vs Inductor in NFB loop?

[potatofarmer]

First, for reference, a redrawn Matamp GT100 schematic - not great, but still more legible than the original scan:


Okay, as you can see, Matamp used an RLC notch filter in the global NFB loop around the power amp for a mid boost when the presence was turned up. There are a few flags warning of stability problems, namely a bypass capacitor across the feedback resistor, a DC coupled cathodyne, and relatively large anode bypass capacitor on the driver tube. Still, I feel like trying this. I don't happen to have any inductors on hand, but I understand they're typically avoided in tube (guitar) amps as they're particularly sensitive to picking up noise.

But if all I need is a notch filter, why not use a bridged-T or twin-T?

This is one of those places where just being a hobbyist on the internet is a big shortcoming, while someone with any level of book-learnin' would be able to answer "well of course it [will/won't] work; obviously."

I'm worried about loop stability, more particularly if there's an effective phase difference between an RLC notch and a bridged-T notch. I took a peak at an online calculator:
http://sim.okawa-denshi.jp/en/RLCbekeisan.htm
http://sim.okawa-denshi.jp/en/TwinTCRkeisan.htm

and putting similar filters in, the phase plots don't look much different. The RLC notch is clearly of higher Q, but since we're talking about a NFB loop with maybe 9-10dB of feedback being applied, either filter is blocking a few hundred Hz.

Granted that this particular implementation of a presence control isn't very popular (and Orange moved away from this circuit when they started building their own amps to one using an inductor between cathode and grid of the driver triode, and eventually just to variable cathode bypassing of the driver) suggests it might be a silly idea or might not work very well after all.

Any ideas or tips?

[Enzo]

Inductors are also avoided because they cost a LOT more than resistors and caps. But an LC is compact and elegant

[potatofarmer]

Sure, there is that. I found what looks like a decent shielded 47mH axial inductor for $1.32 from Mouser - 542-8250-473K-RC - which would probably scare off a large manufacturer. I also hate ordering just one thing from Mouser.

I'll probably wind up trying both ways just for the heck of it.

[Shadrock2]

You should consider the impedance of your circuit and the frequency you want to influence.

For guitar the relevant mid-range frequency is likely around 350 - 400 hz. The inductor and the cap are going to need to have an individual impedance comparable to that of the circuit they are in. That would be measured at 2x the center frequency for the inductor and 1/2 the center frequency for the cap.

The formula for series LCR impedance is given by Z = SQR[R^2 + (2*PI*f*L - 1/(2*PI*f*C))^2]
Sorry about the math syntax but its the best I can do. Hopefully it makes sense.

I considered using a LCR circuit in the cathode section of a triode to provide a switchable mid boost. To have much effect, it would have to have and inductance of around 500mH. This would have an impedance of 3140 ohm at 1 kHz. I haven't actually built this circuit and the numbers are just ballpark to provide a starting point.

The largest shielded inductor I could find from Mouser was 150mH.
A tiny little 47mH inductor is likely to not have much effect unless its a particularly low impedance circuit. (Or higher frequency)
The math for the phase change is pretty much beyond me.

[nickb]

The the RLC, far away from the resonant frequency the phase shift is least. As you increase frequency, the phase lags approaching -90 degrees due to the capacitance then rapidly switches to +90 as the inductor kicks in. Continue raising and phase shift goes back towards zero. All this is happening in the mid band where it's away from the inherent loop phase shifts at the band extremes ( <100Hz, > 15KHz) . Therefore, I don't think you'll see a stability problem.

[Shadrock2]

So I've seen plots where the phase angle switches back and forth at the resonance point. The swings never quite get to 90 degrees.

Nick B., are you saying that phase angle changes of less than 90 degrees are not harmful to amp stability?

[nickb]

So I've seen plots where the phase angle switches back and forth at the resonance point. The swings never quite get to 90 degrees.

Nick B., are you saying that phase angle changes of less than 90 degrees are not harmful to amp stability?

Agreed it will never quite get to 90 degrees either side of resonance - but getting there - nothing is ideal. I'd be surprised if you had stability issues with this kind of phase shift. You need 180 in total which has to come from the power amp. At the LF end coupling caps and transformer are the main issues, at the high end there are a host of capacitances and the output transformer that will do it. Here in the mid-bans where you wanted the resonance the phase shift is least. Still things happens, for $1.50 and a few minutes you'll find out for sure.

[Steve A.]

Getting back to bridged-T circuits I found one in a friend's 1965 Epiphone amp and drew it up for reference. I added it to my EVJ which had the resistor and capacitor values modified to those in the Gibson Les Paul Jr. reissue amp.

I thought that adding the bridged-t circuit made the EVJ sound more like a BF/SF amp which I imagine is why Gibson added the circuit to many of the tweed-like amps they made in the mid-60's (basically the ones with a single tone control or none at all.)

I found thread I posted here on the Bridged-T mod:

Bridged-T Mod for the new LP Jr amp - AMPAGE Archive

Here are two links from initial post:



http://www.blueguitar.org/new/schem/..._lp_jr_amp.pdf

Steve Ahola

P.S. Here is a TSC screen capture using the Big Muff tone control model (I wish I remembered what the 4 traces represented.)

http://www.blueguitar.org/new/pix/ts...d-t-033006.jpg

[Steve A.]

5F2H Schematic with Bridged-T circuit

Bruce Collins had posted a link to the 5F2H amp he designed for Ted Weber in the thread from the Ampage Archive. The URL changed since then so here are the current links:

http://www.tedweber.com/media/kits/5f2h_schem.jpg

http://www.tedweber.com/media/kits/5f2h_layout.jpg

I attached the images here just in case the URL changes again. Here is the audio portion of the schematic shown in-line to facilitate discussion:










Steve Ahola

P.S. Here is a link to the Weber WSE15 output transformer specified in the schematic:

WSE15 Transformer

Here is a link to a $44.10 120v power transformer on the Weber site rated for 100mA for 660v secondary which might be the one in the kit they offer currently:

W022772 Transformer

They also sell an international version of that transformer for $34.65 with 100, 120, 125, 220, 230, 240 volt inputs, 600V and 660V HV taps. (The 660V tap is rated for 80mA. "Note this transformer has wires exiting from both sides. It should be converted to a stand up type by using stand-up bells below (sold separately). Please see pictures and descriptions for mounting info."

W022772INT Transformer

Here is a link to the EU transformer with 120V, 220V & 240V inputs for $50.40:

W022772EU Transformer

Here is a link to the choke referenced in the full schematic ($26.25):

W014684 Choke

[eschertron]

I had to re-draw the filter networks to see that they match what I've come to know as the "big muff" circuit in Duncan's TSC. I hadn't realized this was so common in guitar amps, thought it was a stomp-box thing. Thanks for the links! I'm fond of this style of control, and have more to look at now.

[Zozobra]

Being able to model the response in TSC is super handy. I spent an hour or so playing with it last night working out how to get a 350-400 Hz cut that wasn't too savage.