Akai 707 Preamp Modification

[ruffrecords]

from the web:

http://audiophool.com/Misc/6AU6%20grounded%20plate.pdf

That's the one I was thinking of.

Cheers

ian

 

[untune]

Thanks everyone for your contributions - I got hold of the RD Handbook and have been following up and reading on various subjects. I think I found the schematic you were referring to, I've attached it and I've also tried to incorporate it into the first stage of my pre (on the right.)

I'm just reusing component values but I'm wondering if I have the right idea. From what I've read, this method of triode operation (supressor + plate to ground, screen as anode) will be ideal for a lower signal levels, i.e. the input stage (~100mV signals from mic and attentuated line). Please correct me if the alternative (screen, suppressor and plate strapped) would be better - research has pointed towards that being more suited to slightly larger signal (1V+) and may therefore be ideal for the second 6AU6 triode. Any comments are greatly appreciated

Another couple of questions that have come up... regards to the power supply filtering, I'm sticking to a 20uf reservoir for the rectifier, and I can make the other three (C1, C10 and C11 on the last schematic I posted) either 40uF or 50uF (no 47uF's available) . Or I could make the first smoothing cap after the choke a 40, and the other two further away from the PS 50's. Is there any advantage to either or is it a moot point?

Lastly, something I only just noticed from the original schematic that I missed when redrawing - both of the heater taps on the power transformer (5.7V and 6.3V) run to a point marked 'X' - i.e. are connected to the cathode of the 6BQ5 valve. What is the purpose of this?

Cheers!
Lee

 

[ruffrecords]

I think I found the schematic you were referring to, I've attached it and I've also tried to incorporate it into the first stage of my pre (on the right.)

That's it.

I'm just reusing component values but I'm wondering if I have the right idea. From what I've read, this method of triode operation (supressor + plate to ground, screen as anode) will be ideal for a lower signal levels, i.e. the input stage (~100mV signals from mic and attenuated line). Please correct me if the alternative (screen, suppressor and plate strapped) would be better - research has pointed towards that being more suited to slightly larger signal (1V+) and may therefore be ideal for the second 6AU6 triode. Any comments are greatly appreciated

This method is fine for the first stage. The second stage can tie anode, suppressor and screen grids together and run at a higher current. I seem to remember I used something like 390 ohms in the cathode and 39K in the anode.

Another couple of questions that have come up... regards to the power supply filtering, I'm sticking to a 20uf reservoir for the rectifier, and I can make the other three (C1, C10 and C11 on the last schematic I posted) either 40uF or 50uF (no 47uF's available) . Or I could make the first smoothing cap after the choke a 40, and the other two further away from the PS 50's. Is there any advantage to either or is it a moot point?

Should make little difference. Notcie how the BBC rpeamp runs at a very low current and uses a large HT dropper and smoothing cap for the first stage to minimise HT hum. The first stage is where you really need to keep out the hum

Lastly, something I only just noticed from the original schematic that I missed when redrawing - both of the heater taps on the power transformer (5.7V and 6.3V) run to a point marked 'X' - i.e. are connected to the cathode of the 6BQ5 valve. What is the purpose of this?

Cheers!
Lee

This is a common technique used to elevate the heaters a few volts dc above 0V. This has the effect of reverse biasing the diode formed by the heater element and the cathode which reduces the induction of heater ac voltage onto the cathode.

Cheers

Ian

 

[untune]

This method is fine for the first stage. The second stage can tie anode, suppressor and screen grids together and run at a higher current. I seem to remember I used something like 390 ohms in the cathode and 39K in the anode.

Thankyou again Ian I'm glad I've at least got the right idea for this stage - I removed the screen bypass cap (1uF) and the 1M resistor from HT to screen that were present in the original pentode design. I imagine they are either no longer relevant when operating as a triode - or perhaps I have missed something. Am I right in thinking that the .1uF and the 1M are negative feedback in the BBC design? I'm also a bit curious about how the cathode is wired - 400uF is considerably more than the 25uF I've got and the 10K + 1k in series as opposed to my 1K. Lower gain maybe?

390R and 39K gives me a starting point for the 2nd stage, cheers for those values

Should make little difference. Notice how the BBC preamp runs at a very low current and uses a large HT dropper and smoothing cap for the first stage to minimise HT hum.  The first stage is where you really need to keep out the hum

Noted regarding the power caps - I was curious due to something I read concerning the corner frequency created by the choke and smoothing cap, a 50uF should put it <1Hz if I worked it out correctly. Only problem I might run into is the physical size but I can work around that. In the BBC design, is it the 16uF and 0.25M supplying the screen that you're referring to in reference to keeping the hum down?

This is a common technique used to elevate the heaters a few volts dc above 0V. This has the effect of reverse biasing the diode formed by the heater element and the cathode which reduces the induction of heater ac voltage onto the cathode.

That's what I thought it was for - I'm moving the lamp to the 5.7V tap (although I may end up having no lamp at all) so is it worth repeating this connection to the 6BQ5 with the 6.3V tap, which will supply all the heaters? I've also seen some designs with a 300ohm trim across the heaters which I assume is a method of balancing the hum

Cheers!
Lee

 

[untune]

Hi all, hope everyone is enjoying the festive season.

Few more modifications and a lot of reading but one thing I still can't seem to get my head around is how exactly you calculate how much voltage gets to the plate/screen of each valve from the HT. I'd like to be able to draw the load lines and actually work out and see the operating points of each stage so I can better understand/tweak what's happening. Transformer sec 250-0-250 VAC, then I assume the rectifier and choke must drop some voltage (I read ~20v for 6X4 but couldn't find a figure), but then according to the original schematic, 260VDC at C10, then through a 5K resistor to give 230V, through 50K to give 170V. The original had V1 running at 60V, which I assume was down to the plate resistor. I'm missing something in how the numbers are arrived at. Apologies for what must be a dumb question.

Other modifications: C9 filter cap was connected to V3 screen - now earthed. C6 filter cap added at V2 stage - didn't think more filtering would hurt, although correct me if I'm wrong. Is C1 in the correct place, or should it really sit on the other side of R7? I've left it where it was originally and always assumed that it made no difference.

V1 triode strapped (grounded plate/screen) and V2 triode strapped (screen/suppressor/plate) with Ian's suggested values. V3 triode strapped but not sure it's right yet, and whether it needs R14 (100 ohm?) or not. Also not sure if cathode bypass cap on the EL84 needs to go up - I believe I read something about the power tube cathode circuit having a lower impedance and maybe needing a bigger cap.

Also not really explored the concept of negative feedback and how it might figure into things, but I'll get to that at a later stage.

Cheers
Lee

 

[ruffrecords]

dc load lines are quite easy. Using the first stage as an example, you have a 250K plate load and the HT voltage is 170V. When the tube is not conducting there is no current flow and the plate will rise to the HT volatile of 170V. So you can mark the point 0mA and 170V on the plate curves. If the tube was a short circuit, the plate voltage would be 0V and the plate current would be 170/250 mA which ia about 0.68mA. You can mark the point 0.68mA and 0V on the plate curves. Join these two points together with a straight line and you have your dc load line.

Now choose an operating point such that the nominal plate voltage is about half the supply voltage. Find where this voltage intersects the load line and that will be the plate current you want.  From the plate curves for each value of grid voltage, estimate the grid voltage needed (sometime I pick an operating point on a grid voltage curve that gives an operating point close to where I want to be just to simplify things). You now know the retired plate current and grid voltage to make the tube operate where you want it. So now you calculate the cathode resistor as grid volts over plate current and you are done.

You will often find, as in this case, that the dc load line occupies a tiny corner of the the tube curves and obtaining accurate grid voltage estimates is difficult.

Cheers

Ian

 

[untune]

Thanks Ian, that's made the process much easier! The thing that was really confusing me is how those original HT voltages are arrived at - I have take the 170V, the 230v, the 260v from the oiginal schematic as i have not changed R12 or R15 - I assume these are responsible for dropping the voltage. I'm confused as to what happens to that original 250-0-250 coming from the transformer to get to those DC voltages. I can't seem to find the proper calculation regarding the rectifier, the choke, the resistors etc...

For example - I believe PRR suggested earlier in the thread that a resistance loaded EL84 would be a good option, and 250v was mentioned. If the HT is at 260VDC after the choke, then has to pass through the output transformer pri before it gets to the plate of the EL84/ - will this affect the voltage and if so, how do I work out by how much?

Same with the V2 stage with the values you kindly provided (I managed to track down old threads and you were spot on with 39K/390R - almost ten years ago, good memory ) and I believe your HT was 250V - but again my HT at the same point is 170V. I'm just a bit lost on how I figure out how to adjust them.

Cheers
Lee

 

[ruffrecords]

There are standard formulae for working out the basic dc voltage from a power supply. One example is here:

http://www.sowter.co.uk/rectifier-transformer-calculation.php

Yours is a bit more complex because it has a capacitor input followed by a choke.

You generally work out HT voltages and current from the input to the output. You choose the HT voltage and operating current of the first stage and second stage. The first stage gets fed from the HT supply of the second via a series resistor and a decoupling capacitor. The value of the resistor is the difference in HT voltage divided by the first stage current.

The second stage is fed from the third stage HT via another series resistor and  decoupling capacitor. This time the resistor value is the difference between the third stage and second stage HT voltages divided by the sum of the first and second stage current.

And so on.

Cheers

Ian

 

[untune]

Thanks for the explanation Ian... It's sinking in slowly but the numbers are melting my brain ;D

I have to get some components ordered by tomorrow so I can get them cheaper, quick question regarding values. I can only get hold of a 22uf cap for the reservoir - I needed a 20 which is already above the recommended but given tolerances etc is it safe to assume it won't hurt the rectifier? I've read of sone people using 40 and 50uf with the 6x4 with no issues but these caps aren't cheap, I'd like to be sure

Same with a tape recorder I need to fix - need to replace a 5000uf/18v but can only get 4700uf/63v - it an acceptable replacement?

Cheers

 

[ruffrecords]

Most electrolytic capacitors have a 20% tolerance anyway so you will probably be OK in both cases.

Cheers

Ian

 

[untune]

Thanks for that! I've plotted the 2nd stage triode at least using your values and think everything is starting to make sense.  Quick question regarding resistors in a circuit like this - what wattage is ideal? I'm unsure if you can get away with lower wattage ones on cathodes since the voltages are low, but are they dissipating power?

 

[ruffrecords]

Quarter watt resistors will be fine for cathodes. The power dissipated is simply the current through them times the voltage across them. Most cathode resistors have less than 10V across them and less then 10mA running through them so they dissipate a lot less than 100mW. In preamps, plate resistors typically have about 100V across them and carry a couple of  mA so they dissipate typically around 200mW so it is probably better to use 1W types for them.

Cheers

Ian

 

[untune]

Thanks Ian, and a happy new year to you!

Sounds like my best bet would be to stick to 1 watters.

I'm getting near to a point where I can start putting something together. Am I right in thinking that the 260v at the cap/choke is the HT I have to work with? It seems that stage 3 (EL84) and stage 2 (6AU6T) can both be run at 250v on the plate. Going off the original Akai schematic, the EL84 is already seeing 250v. If stage 2 uses your values of 39K/390R then running it at 1.5v bias gives plate voltage of ~100v and somewhere around 4mA.

Stage 1 I've been working out based on also using 250v,but perhaps I need to drop this down with a resistor - not sure if there's an advantage to lowering it here. Tube life perhaps?

With stage 1 as a grounded plate/screen triode, 250K on the plate and 8K on the cathode, 4v bias sits at 120v plate voltage and 0.5mA current.  I've lost it amongst my notes but I seem to remember <1mA being important here.

If all this sounds about right, then it's just the output to figure out!

Cheers
Lee

 

[ruffrecords]

It sounds good as a first stab. The only thing to think about is the 'anode' voltage for the first stage. With the screen grid as anode set up you have you need to take care not to exceed the screen grid dissipation which is just the anode current times the anode to cathode voltage. As this is somewhat lower than if you had used the anode itself it makes sense to consider lowering the anode voltage to keep this in spec. As the signal is still small at the output of this stage you do not have to worry too much about the voltage range over which the anode can swing. It also allows you to include some additional decoupling so the HT noise/hum is minimised because the power supply rejection ration of a simple stage like this is poor.

Cheers

Ian

 

[untune]

Dissipation is marked as 0.65w for this type of triode setup - I've kept well away from the line when working it out so things should be fine. I'll update the schematic with changes when I get a minute... Back to work tomorrow

Speaking of the decoupling, on every schematic I've drawn up I've kept cap C1 in the same place I originally found it - to ground after the plate resistor of the first stage.  Shouldn't it sit prior to the plate resistor or does it not matter in that instance? I've noticed that in some later models of the same recorder, that cap goes to ground between 1st and 2nd stage

Cheers
Lee

 

[ruffrecords]

Not sure what you mean by prior to the first stage anode resistor - wouldn't that be the anode itself?

Cheers

ian

 

[untune]

I'm an idiot - I'm looking at the schematic a bit too literally and just realised the cap won't care which 'side' it's on  : Sorry!

A more logical question - PRR spoke of a resistance loaded triode EL84. It's my understanding that the OT acts as a plate load in a regular SE output stage, but how do I apply that resistance - have I got the right idea with a resistor in series with anode and screen connected? There unfortunately doesn't seem to be alot of information out there on it

Cheers!
Lee

 

[ruffrecords]

A more logical question - PRR spoke of a resistance loaded triode EL84. It's my understanding that the OT acts as a plate load in a regular SE output stage, but how do I apply that resistance - have I got the right idea with a resistor in series with anode and screen connected? There unfortunately doesn't seem to be alot of information out there on it

Cheers!
Lee

Treat it just like the other triode stages. The EL84 data sheet provides curves for triode operation:

http://www.r-type.org/pdfs/el84.pdf

Your HT is around 250V. As this is an output stage it makes sense to arrange the anode voltage to be about half the supply voltage. The data sheet says you can run up to 34mA through it as a triode but this is far more than we need. 25mA should be plenty. This means the anode load resistor is 5K (5K x 25mA = 125V). YOu can draw this load line as before and work out the grid volts for 125V on the anode and then calculate the value of the cathode resistor.

Cheers

Ian

 

[untune]

Thanks Ian, makes perfect sense. I think PRR suggested 2K7 and 30mA, 167R on the cathode, so I'll draw some lines tonight and see what everything looks like. Will this also affect the choice of output transformer? I believe I can get away with ungapped in this arrangement, but I found some Sowters that were designed for single ended line output and they were £150+!

Speaking of things making sense, dropping the HT voltage with resistors based on the current draw of each stage finally clicked last night, but it got me thinking - is there an advantage? If I have the EL84 and the 2nd stage both using ~250V HT at the anode, I would need to drop a lot of voltage over one resistor to get the first stage down to, let's say 180v. If the first stage is only drawing maybe ~0.5mA, then (70v/0.5mA) that would be a 140K dropping resistor? Or the other option is to maybe drop stage 2 from 250V by about 30V (maybe ~4mA so 7K5?) and then that down again by another 50V (50/.05) = 100K. There shouldn't be a negative effect on the corner frequency with 47uF smoothing caps, so perhaps bigger resistors = more noise considering the power they're dissipating? Might have it all wrong but I'm not sure if there are rules to follow here

Cheers
Lee

 

[ruffrecords]

The exact anode load value is not critical. At 25 to 30mA the anode resistance is about 2K. For low distortion it is preferable that the plate resistance in several times this value. 5K will probably be OK. The output impedance of this stage is approximately the anode resistance in parallel with the anode load resistor which would be about  1.4K with a 5K anode resistor. You need to transform this to a value considerably below 600 ohms. If you use a 4:1 transformer the output impedance will be transformed to one 16th of its initial value so it should be below 100 ohms which is about right. A 9600:600 such a a Crnhill VTB2291 will do - available from Audiomaintenance for just over £20.

Splitting a single RC filter into a series of RC filters provides more ripple attenuation than a single RC with the same total resistance and capacitance. This was well known to early valve equipment designers. If you look in the power folder of the DIY tab of my web site you will find an article from the October 1949 issue of Wireless World by Mr. Scroggie that into some detail about it.

http://www.customtubeconsoles.com/diy

Cheers

Ian