MOSFET source follower for output stage of AB763 blackface

[martin manning]

In the schematic I posted I show 100uF 250V caps for the bipolar supply. For bleeder resistors I think 180K 1/2W would be good.

[stephenl]

interesting, informative thread. I'm anxious to see how this works out for you SSB. I might have to try this one out.

[John_P_WI]

martin, yes thank you for the verification of the 12V Zener across the DRAIN and SOURCE.

Shouldn't it be Gate to Source??

[SixStringBender]

Martin, thank you. I was asking my wife what was leading to the 160 and -160V supply in your schematic, but she couldn't figure it out. I found a bipolar power supply schematic doing a google search last night that was using a CT transformer so I had her run me through it. You should have heard her describing the AC waves from the secondaries. Her words, "hey it looks like a hand." A hand I asked. "Yeah, make a fist," she said. "It looks like your knuckles with a line in the middle of the middle two knuckles." lol HAHAHA

Here is the schematic of the bipolar supply using a center tapped transformer. Figure 8.
http://www.bristolwatch.com/ele/power_supplies.htm

I figured out the center tap was the line in the middle and the knuckles were the AC. lol I had her look up an AC wave just to be sure.

From there I found a description, in text, of how to build a bridge rectifier. So I visualized that in mind (I could see good enough to read until I was about 19 and could read a little until 23 from the T.V. screen). Then I found R.G.'s description of how the smoothing caps go to ground in series. That was my eureka moment. I was like George Lopez. I got this!

An informing thread? Yes sir it is. It has to be for me to figure it out. lol

[SixStringBender]

Yes it is gate and source. I don't know why I typed that. I gotta fix that too. Thanks for pointing that out so someone won't make that mistake.

I've learned this before from the valve wizard. I think it is done to prevent an arc at power up. There was even a resistor in series with the zener in that article to prevent switching noise from the zener.

martin or R.G. have any info on that resistor to prevent switching noise?

[martin manning]

Here is the schematic of the bipolar supply using a center tapped transformer. Figure 8.
http://www.bristolwatch.com/ele/power_supplies.htm

Figure 8 at this link is electrically the same circuit that I drew, just arranged differently.

I don't think you will have any problem with arcing as described by Blencowe (unlike a tube the FET doesn't have to warm up to begin conducting), and you should have plenty of headroom so no switching noise.

BTW, the commonly available 200V rating is high enough for the power supply caps.

[SixStringBender]

Sounds good, thanks martin.

Studying the bias circuit of the Fender schematic confused me. Luckily I found a great description of a Super Reverb bias supply. I interpreted it to the DRRI as follows.

The below is edited with the correct description of the DRRI. The Super Reverb had a small difference that caused me to make a mistake (original SR description most likely). The smoothing capacitor is on a different pot lug in the DRRI than it is in the description of the Super Reverb.

Stock DRRI bias supply:
R58's center lug (wiper) goes to C29's negative lead. Wiper also goes to R56 and R57.
C29's positive lead to ground. C29 is 22uF 63V polar electrolytic.
Negative bias supply voltage goes to R58's input lug.
R58's other outer lug goes to R59 (10K 1/2W). Other end of R59 goes to ground.

Dual bias supply:
Disconnect R57 from the wiper of R58.
Connect R57, and the negative lead of a new 22uF 63V electrolytic cap, to the new 10K bias pot's wiper.
Connect the positive lead of the new 22uF 63V cap to ground.
Connect input lug of the new 10K bias pot to input lug of R58 to share the negative bias supply.
Connect the other outer lug of the new 10K bias pot to a 10K 1/2W resistor and connect the other end of the 10K 1/2W resistor to ground.

The above is edited with the correct description.

The below is edited to add the cap to the description.
When it comes time to do this I have a feeling there will be a short trace between R56 and R57. One or the other will go up to the solder pad that goes to the wiper of R58. Witch ever has the short trace going to the other will be the one that gets disconnected when we cut the short trace that joins the two. So it may end up being R56 that goes to R57, then R57 to the wiper of R58. If so, we'll put the wiper of the new pot, and the negative lead of the new 22uF 63V cap, to R56 and leave R57 in tact. I believe that is okay if it happens that way.

If anyone attempts this in the future on a PCB amp always ohm across the cut trace to make sure there is no longer any continuity. In this case from R56 to R57.

[martin manning]

Dual bias supply:

Disconnect R57 from the wiper of R58.
Connect R57 to the new 10K bias pot's wiper.

>>>Connect a new 22uF 63V electrolytic from the new 10k pot's wiper to ground. This cap replicates the function of the existing bias supply's C29 in the new bias supply. [MPM]<<<

Connect input lug of the new 10K bias pot to input lug of R58 to share the negative bias supply.
Connect the other outer lug of the new 10K bias pot to a 10K 1/2W resistor. Then connect the other end of the 10K 1/2W resistor to ground.

I don't think the bias pots can share R59.

Looks good except see my added comment in the quoted text above about a new 22uF electrolytic (as shown in my schematic posted above), and you are correct about the supplies not sharing R59. If they did they would interact.

[harleyboy2112]

Here's something I can across about ten years ago or more....

It was around the time RG wrote the Mosfet Folllies and I think he helped this fellow get this circuit working...

It has an alternative bias supply that's not as refined as Martin's but may work for some who want to use the existing supply....

TT

I've been following this thread because it's something I've been wanting to try. What are the thoughts on tictac's schematic?
Is it a viable option for the mosfets?
It seems like it would be easier to use a tube instead of building a separate power supply for the mosfets (if you have the space and required transformer power).

[R.G.]

It seems like it would be easier to use a tube instead of building a separate power supply for the mosfets (if you have the space and required transformer power).

Using a tube instead of a MOSFET makes the DC levels problem worse, not better.

You have to have a power supply of some kind that will let the MOSFET or tube drag the output tube down to the same DC voltage it would normally have if there was no buffer. In order to do that, either a MOSFET or a tube needs a source/cathode resistor connected to an even lower voltage, by some small amount at least.

A normal MOSFET's gate is a few volts more positive than its source. A tube's grid is one to several volts more negative than its cathode. The follower tube's grid then has to be held even more negative than a MOSFET's gate to get the cathode to be at the same voltage to drive the output tube's grid. So there has to be a bigger negative supply for a tube to work from in the same application.

[SixStringBender]

harleyboy, plus a tube needs heater supply. In the case of the AB763 you'd have to lose the tremolo tube too. So you'll lose that function. And as R.G. said, you still have to build the bipolar supply.

If you do use a tube I would use the 12AU7 instead of the 12AX7.

martin, thanks. Maybe the Fender DRRI schematic didn't have me as confused as I thought. I thought the wiper of R58 was going to C29 on the schematic, but I hadn't checked the board to verify. In the article that described the Super Reverb's bias supply it lead me to believe the electrolytic cap (C29 on the DRRI) is going to ground from the input lug.

Here is the article.
http://el34world.com/charts/fenderservice5.htm

I guess the cap is going to ground differently in the DRRI. I'm glad you caught that. I'll make note of it. I was assuming R58's input lug had the cap from it to ground like the Super Reverb description.

So the new 22uF 63V electrolytic cap's negative lead to wiper of new pot and positive lead to ground.

I will edit my above post with the correct bias supply info for stock and dual bias. I want the correct info to be provided for others. So for future readers, if you see "edited with correct description" at the beginning and end of the description it has been corrected. Refering to my last post.

I have a question about the pots. Could I use two 25K multi turn trim pots for more bias range when I may install KT66 or 6L6GC in the future when I need more volume? Or do the 10K resistors (R59 and the new 10K to the other pot) have to be replaced with 27K when that time comes (using the 10K bias pots). It seems like the pots are just in series with the 10K resistors to ground, so I would think 25K multiturn trim pots would work.

[SixStringBender]

In addition to the 25K bias pot question I have another.

I found the equal power transformer to the SMD one in a PCB version. I know I can use it, but can I use this one if I choose to chassis mount it? It is the lowest output chassis mount power tranny I can find. I have searched for hours. It is twice the output of the SMD and PCB mount trannies. And it is in stock. 48 of them available!

It may even fit inside the chassis, but I'll probably mount it outside in the air. I just think a chassis mount tranny with leads will be much easier and quicker to work with.

http://www.mouser.com/ProductDetail/Tri ... MqnA%3d%3d
Mouser part 553-VPT230-110
$17.55

edit: The Hammond 185C230 is the same spec as the VPT230-110 Triad.
Both are chassis mount, same 230V output and same 25VA 0.11A output. Their dimmensions are different.

The 185C230 is in stock too.

[martin manning]

Yes, you can sub a 25k pot for the 10k and that will have more range but the maximum negative voltage won't be much lower. The schematic says the raw bias voltage, at the top of the pot, save for a 22 ohm resistor, is -49V.

The transformers you linked are quite a bit bigger than the flat-pack I suggested, and have double the VA rating. A flat-pack can be chassis-mounted upside-down, with the pins sticking up, and then you can easily solder leads to them. If you can find room inside the chassis, this would be a good solution.

You could even put a little board on top of the transformer to hold the two caps and the rectifier, making the footprint for the whole power supply 2" x 2.5", and it would be less than 2" high. I see that the flat-pack transformer is back-ordered, but the estimated ship date is only a week off (4/23).

[Kagliostro]

About the transformer, also Hammond has flat-pack units

http://www.hammondmfg.com/pdf/229Insert.pdf

Franco

[John_P_WI]

SSB, maybe you could use this for an idea:

http://www.londonpower.com/all-products ... supply-kit