Discussion in 'Fisher' started by dcgillespie, Mar 29, 2013.
Yep both channels. As metioned before I suspect about the DC voltage that feeds the EFB.
Hi Luis -- Sorry you're have trouble. If the AC supply is low, then all B+ voltages will be low. As for the screen voltage specifically, the EFB screen regulator will cause -- and hold -- the screen voltage to always be about 73% of of the main B+ voltage value -- whatever that value is, under whatever the operating condition.
In my own unit, the main B+ runs about 430 volts under quiescent conditions, producing about 312 volts at the screen grids. As this voltage goes up or down either due to changes in line voltage, or changes in available power supply voltage due to power demands placed on the amplifiers, the screen voltage will always remain at about 73% of whatever that main B+ value is.
Of course, the EFB control grid regulator performs the same basic way, causing the negative bias voltage applied to the control grids to go more negative as B+ increases, or be reduced (less negative) as the B+ is reduced.
This all happens automatically, meaning that instead of a low line voltage being the issue of low quiescent current, it is much more likely due to the fact that the particular set of output tubes you are using requires a lower than nominal bias voltage to achieve the target current draw. If that is the case, the 39K resistor in the EFB control grid regulator circuit can be reduced to 36K, or even 33K to bring the adjustment range of the bias control in range to the particular set of tubes you are using.
Finally, one sure fire way to find out what's going on would be to run some power output tests. With the quiescent B+ and current levels specified, my unit has no problems easily producing about 23 watts per channel, with both channels driven, from 30 Hz to 20 kHz. Mid band distortion should be on the order of about .15% or less under this condition.
Let us know what you find, and I hope this helps!
Thanks for your support!:thmbsp:
In the best scenario I can get from the mains 403VDC.
But now you are clarifying something:
Whatever the +B in the output of the EFB must have 73 % of it, but in my EFB board something maybe is defective because when I'm increasing the cathode voltage, the screen voltage is going down at the point where with the resistor changed (I did this before and now is 33k) I can reach only 16 mA per tube and the screen voltage is in 130VDC
Dave with no output tubes how many volts would be the cathode voltage and screen voltage??? FYI a week ago a 10 ohm resistor 1/4 watt was found open, is possible that the EFB board ( with some defective) is responsible in open the current bias resistor?
The tubes I'm using are a GE top getter
Hi Luis -- To have a screen voltage of just 130 vdc implies that the main B+ is just 180 volts -- which is impossible to believe unless something is grossly wrong with the power supply.
With no output tubes installed, the cathode voltage would be 0, since there would be no tubes installed to draw any current through them. And, with no output tubes installed, the screen voltage would be -- again -- about 73% of whatever the main B+ rose to with no output tubes installed.
It sounds like either the 1.8K Drain resistor, the 100K Gate resistor, or the mosfet itself in the EFB screen regulator as failed. If so, it will be the first instance I am aware of where any of these parts have simply failed.
Since you had a cathode resistor fail, it certainly implies that one of the tubes went south. You might notice that I did not include Screen Stability Resistors in the modified output stage. That's because they simply are not needed to control any potential output arcing events in this design, because the design is not one that promotes output tube arcing. However, that still doesn't mean that a tube -- of its own accord -- can't be defective and produce its own arc.
If you want to install them for extra protection, a 1/8 watt device is the appropriate wattage rating that will still allow continuous full power output if desired, but will open on overload.
Still, the specified mosfet is a 9 amp device with an electrolytic cap tied directly to its Source terminal, and should therefore be able to take any such overload in relative stride. Unless the device was previously damaged, I have not been able to cause mosfet failure in my development tests.
I'd zero in on this area, where I'm sure you'll find the problem. Let us know what you find!
Just a note. I have seen electrostatic handling issues partly damage Mos-Fet devices, yet still functions for a limited time. Then complete failure of the device occurs later.
The Mosfet was defective(James you was right!), also the voltage divider that feeds the gate in the 100 k resistor after change the mosfet got open, so after replace the Mosfet and the 100k resistor with the 33kohms in the bias circuit on the ibam the ouput was too high (around 25V ), so I had placed again the 47k resistor described by Dave in their drawings and now the 400 is sounding very well again!!!
Thanks for the support!!
Glad you got it singing again!
A bump for a good thread!!!!
These links don't seem to work now.
This is one of the artifacts from the changeover that is being addressed. It may take a couple of months for all the data to be changed over, adjusted, cleaned up, etc.
In the mean time, I would PM Eduarsan and ask him to email the files to you.
Thanks, l'll do that.
Here there are the last drawings!!!
Many thanks, Luis!
Dave, Would an LME49720NA be a suitable chip replacement or upgrade to the 353? Thanks!
In theory, there is no doubt that better ICs can be used for the buffer amplifier. The question is if you'll ever realize the improvement. It would be one thing if the chip was providing actual gain as a product of the circuit it is installed in. But in the buffer circuit, it is operating with 100% NFB producing zero voltage gain, so pretty much, the output is a exact replica of the input, but at a lower impedance. In other words, distortion produced by any compatible IC is going to be difficult to even measure, let alone hear.
I have found that when it comes to ICs, better isn't always better. Oh, it is in theory and on paper. But I've seen plenty of high performance chips that become so cantankerous to work with (i.e stabilize), that the collateral damage from using them actually makes their use less desirable. For the purpose of the circuit, the specified device is going to give you 99.99% of the desired end result. I know it can be fun chasing that last .1% -- but it doesn't usually come easily. That being said, the 353 is a very fine upgrade, and should work perfectly. As to whether you can tell it's installed is another matter entirely.......
It's what I have on hand. Would it be ok to use in this circuit with the other components listed, Or should I just order the 353? Thanks for your assistance!
It should work just fine, since the application notes says it remains stable in a unity gain configuration. With such an extended response however, make sure that your circuit layout is well thought out, and that the supply bypass cap is attached right at the supply terminals of the chip socket.
Will do. Thanks again Dave!
Is there anybody on here willing to sell kit version of this with everything needed to do this mod. I enjoy doing the work but finding everything from different sources is a pain.
Ask to Larry probably he can assist you!!!
Separate names with a comma.