Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: Papa Jim on March 22, 2019, 07:26:30 am
-
If cost is not an issue because you already have the iron, and the space is there any reason not to do a 4 diode bridged full wave P.S. ?
-
The voltage needs to be appropriate for a bridge.
-
Some want the sag from a rectifier tube.
-
The voltage needs to be appropriate for a bridge.
Sluckey. Meaning the final B+ Voltages because it will be higher than if you used the same P.T. with a half wave, not that the diodes care as long as they are rated correctly?
Some want the sag from a rectifier tube.
Okay Willabe. I need to google what the sag does for the tone.
-
Half wave rectifiers are rarely used for B+ supplies except in very cheap, bottom of the barrel type amps. Half wave works well for bias supplies.
Why don't you tell us which transformer you have, or at least give us the secondary voltage. Also tell us what ideas you have in mind.
-
Sure will. I just have a couple things on the schematic to finish and I will post the whole amp Sluckey.
-
Here's my schematic. First time posting a Ltspice schem. Let me know if you can't see it.
-
it's typical to keep pre ground all by itself away from the dirty grounds
-
This is the circuit you mentioned yesterday in another thread...
Don't know if this helps any but I use a 250-0-250 ( Hammond 269JX) 50va 69ma using the CT and one end with a 4 diode bridged full wave arrangement, for higher voltage at the expense of less current though. I'm getting real close to both the B+ voltages and 12AX7 plate 150V shown on the schematic of a similar to champ princeton amp. I did use a choke, and higher value caps though.
Using a bridge on half the secondary will produce exactly the same B+ as using a conventional 2 diode rectifier with the entire secondary winding. There is no advantage to using the bridge and only half the secondary winding. In fact, there may be issues with doing that. As shown in your schematic the upper half of the secondary is working full time and having to produce all the current needed by the amp. Meanwhile, the lower half is resting full time. This may cause heating problems with the secondary, but I don't know for sure.
Now using a conventional two diode full wave rectifier (as that PT was designed) will result in the upper half of the secondary working on one half cycle of the AC and the lower half of the secondary working on the other half cycle of the AC. Each half of the winding is working half the time and resting half the time. IOW, the entire winding is being utilized, spreading the work evenly. This should also spread the heat evenly across the entire winding. Just sounds more logical to me.
Obviously the circuit you have drawn will work but that doesn't make it a good idea. It's not something I would do.
-
Ok. I was only thinking about how to get a close to final voltage and by 250 x 1.414 and missed the whole current thing, and I really didn't know I could get there and avoid the heat and current issue. Thanks Sluckey. I will change that for sure.
-
If you ground both ends of the shielded cable it will form a ground loop.
Just ground 1 end of the shield.
It's usually grounded at the quite end, in this, the input jack.
-
it's typical to keep pre ground all by itself away from the dirty grounds
Hey Shooter, I'm confused about star single point ground then. What am I not seeing. The guy that had me just build a Deluxe reverb had me do it this way and it was very quiet. That is the only reason I copied how he had me do it. :help:
-
If you ground both ends of the shielded cable it will form a ground loop.
Just ground 1 end of the shield.
It's usually grounded at the quite end, in this, the input jack.
Willabe, I think it is grounded at one end because of the isolated jack sleeve/washers. (maybe blurry writing on my schematic) I could move it to the quiet end of the cable though. Take a look again. I definitely want to get that right. :icon_biggrin:
-
Willabe, I think it is grounded at one end because of the isolated jack sleeve/washers. (maybe blurry writing on my schematic)
Yeah, I can't read it.
If it is an isolated jack, then then it's fine as drawn.
-
it's typical to keep pre ground all by itself away from the dirty grounds
Hey Shooter, I'm confused about star single point ground then. What am I not seeing.
Read these on grounding, if you haven't yet.
http://www.aikenamps.com/index.php/grounding
http://www.valvewizard.co.uk/Grounding.html
-
Is this better, for the gnds. But I haven't done the tranny change yet sluckey
-
there are 2 conventional ground symbols, one looks kinda like a rake - chassis ground, the other looks like an upside down triangle - signal ground.
chassis grounds are "typically" the noisy one's, often with noise levels = to your input signal, by forcing signal ground to "use" the same physical point, you might
be inducing that noise onto the signal, causing the dreaded 2 week hunt for HUMM.
-
I usually never draw in the wires from the stages to my ground point. Usually just put a label like you did (C*4)....etc. but not a ground symbol. Then I do a matching label at the ground point.
I'm just going by your drawing and so was Shooter.
I use a star ground and buss. Kevin O'Connor calls it 'galactic' grounding, separate ground stars that are then wired to a ground buss wire and only 1 chassis ground connection. It's a wired ground scheme, instead of random chassis grounds.
What ever tube/tubes a B+ filter cap feeds, those circuit grounds go to that filter caps ground, in a star. Then you run a ground buss wire to the next ground star.
Aiken and Merlin talk about the same thing.
Shooter was talking about separating the power tube ground star (high current) from any/all preamp tube ground stars (low current) so the higher currents don't modulate the low current/low signal tubes and get amplified.
The noisiest ground is the PT's CT to the 1st filter cap. They should be wired to each other directly and then you run a buss wire to the power tube ground star. That way the filter cap charging/discharging currents flow in a small loop and will not disturb/modulate anything else.
It's in the links. Aiken and Merlin both talk about this clearly.
You need to read Aikens grounding again and read the other link, Merlins.
I think for most people, myself included, you have to read it and read it again, think about it for a day or 2, and then read it again, until you really understand it.
-
okay I will do that and hopefully have it straight. Maybe. :laugh: I hope.
-
If you ground both ends of the shielded cable it will form a ground loop.
Just ground 1 end of the shield.
Guess what happens if you only ground one end of a guitar cable? Or mic cable.
-
> Here's my schematic.
"Single point" or "star" ground is often misunderstood.
And IMHO, not what a guitar amp wants.
{as others say} Your plan has a BIG problem. The path from the rectifier to the first cap always has BIG 100/120Hz CRAP on it. Here you have routed all that crap THROUGH the common path for all your precious signals. This will buzz!!
The loop through PT, rectifier, and first cap must be *small and short*. We must not make any connection ahead of the first cap to the rest of the amplifier. Yes, sometimes a few inches of wire common to crap and amplifier is not enough to be a problem; but sometimes it is and causes much angst.
Treat the PT-Rect-Cap as its own thing, like it was a 300V battery.
From there, you can sometimes return everything to the - end of the first cap. And chassis-ground wherever. However IMHO Ground follows Signal. Power goes where big power is needed. So _I_ run the - end of the 1st cap to the power stage. Then a ground bus following the signal path (only backward).
IMHO, if you use metal jacks, your ground bus lands at the input jack(s) and goes to chassis there. If you use plastic jacks, connection to chassis is normally best but the location is not real critical.
There ARE other ways to skin this cat, "One point star" can be done well. For for a beginner, tight 1st-cap loop and then bus is probably the easiest to get right.
-
Agree that if you have a CT winding you want the 2-diode form of rectifier.
Your NFB at V2 will not work with a cathode cap there.
I do not see a ground reference on your heater circuit.
-
I ALWAYS use a non isolated input jack and carefully scrape off the metal of the chassis before installing it with a star washer on the inside of the chassis to ensure a solid contact.
This is my preamp ground bus where the pots gnds, cathode resistors and caps, grid leaks res. and shields from signal cables are connected to.
-
You can use isolated input jack as well but you have to drill a hole in the chassis and do the metal scraping work just the same with a connector and piece of wire to the input jack ground terminal.
-
Oh wow! I go take a nap and all hell breaks loose. :l2: :l2: I'll take a look guy's and respond soon. :icon_biggrin:
-
If you ground both ends of the shielded cable it will form a ground loop.
Just ground 1 end of the shield.
Guess what happens if you only ground one end of a guitar cable? Or mic cable.
You know what I meant. :laugh: I've seen you post the same thing.
How about; Just ground 1 end of the shield on the chassis' internal ac signal shielded cable.
-
I ALWAYS use a non isolated input jack
Here you have routed all that crap THROUGH the common path for all your precious signals. This will buzz!!
you have to read it and read it again, think about it for a day or 2, and then read it again, until you really understand it.
highlight reel :icon_biggrin:
-
Agree that if you have a CT winding you want the 2-diode form of rectifier.
Your NFB at V2 will not work with a cathode cap there.
I do not see a ground reference on your heater circuit.
Oh man thanks for finding that NFB problem and the no artificial ground on the filaments. I will fix all of these things and re-post my schematic for round 2. :laugh:
-
Ok Guys Round 2. Fire away and I will fix. :icon_biggrin: Do I need to figure out how to send as a .gif file Until them maybe PRR would be so kind to post again for me, and if it is correct no people and trucks. :l2:
-
Nothing wrong with png files. But your pic is too small to see any detail.
-
Thanks Sluckey. Take a look now. From my end the image is a lot better now.
-
I still can't read component values or voltages. Can you?
Need a filter cap on both sides of the choke.
-
I still can't read component values or voltages.
Nope, too small.
-
Not real good. I will work on that guys. The problem is I am screen capturing an LTspice Schematic and the captured image is not great. If anyone uses LTspice and can tell me how to do it better let me know. Thanks.
-
Click on help.
-
Screen capture is full of round-off errors which make sweet sharp black/white (blue/grey?) line drawings blurry.
The first thing I did was to whomp to black/white so I could stand to look at it. Even then most of the small print was too small and blurry to read. Not getting paid to squint. I just put some crayon and clipart on.
PNG is a sophisticated image type, with ability to do smooth skin-tone gradients on your girlie-pictures. It "can" do crisp line-drawing, but the defaults for general screen-grab probably opt for "smooth" rather than "crisp". I use GIF because it is more than good enough for line-drawing, and will NOT fuzz-up details trying to "make nice". (GIF originated when 16-color was gee-wizz and smooth nudies were only in our dreams.)
I would compact the layout somewhat, and if possible bump-up the font size. I did that permanently in my sim program, from 0.08" to 0.1". (This program is so old it does not do Points.)
Yes, there is usually a way to Export To Decent Picture. Maybe try what Sluckey says?
-
Menu Bar - Tools > Copy To Clipboard.
Open Paint > click paste button.
save as a PNG or GiF - GIF will have a lower resolution.
attached is an example i created using the above - the file was provided with the LT-Spice distribution.
--pete
-
Test for better picture
Hope these are readable. If not I will try HELP. :icon_biggrin:
Sluckey please double check for my first ps cap. I know it looks weird drawn in a cap can like I did. Also guys the artificial filaments gnd. Where physically is the best place to attach to that buss? Closer to the input jack or the OT end of it.
Can't wait to get into the guts of my amp and fix all of these things.
P.S. Sluckey I did go back and remove my incorrect post about the PT in that "tale of two tweeds" thread, so as not to lead some poor soul astray. Thanks for jogging my memory on that one. :icon_biggrin:
-
Much better pics. I bet you can find a way to hide those grid dots too. :icon_biggrin:
Fender connected those filament resistors to the indicator lamp socket on the front panel but you can mount them anywhere along the filament string, even at the very end. Another trick is to mount them on the 6V6 socket. One resistor between pins 2 and 8. The other resistor between pins 7 and 8. This connects the resistors to the 6V6 cathode for an elevated dcv reference instead of ground reference.
You NEED a filter cap at B+2 node.
-
This should be even better without grid dots and with Sluckeys find of no B+2 Filter cap corrected
-
There's a more conventional way to draw that cap can that makes it easier on the eyes. Here's an example...
http://sluckeyamps.com/sunn/sunn_sceptre_1971.pdf
-
Hows this.
-
Thanks so much for all the great help guys. I am starting today to fix all those mistakes on this amp. I know it will sound much better when I do. :icon_biggrin:
-
Ok. Fixes complete. Ready to power it up. Whats the protocol.
Without tubes power it up through a series connected light bulb?
Take B+ High Voltage measurements for ballpark values.?
Power off. Install tubes and repeat.
Play through it if all looks good.
-
before I even apply power I hook up a meter for ohms, connect + lead to the last PS tap, put the - lead to a tranny bolt and watch. you should see big K ohms "charging" up to many many k ohms.
I call it smoke prevention test :icon_biggrin:
If you have a nice 5K 50W R laying around, clip it where the meter was, then clip meter across the R, set for VDC. Should give you a pretty good "load" to test the PS without tubes. Once that's complete, add tubes. It's a SE, so have the meter on the cathode R, set for VDC. I set up my second meter to the PS for the PA tube. Power up, as soon as filaments stabilize, ~ 30sec to 2min, grab the 2 readings, power down and do math.
-
will do
-
Tube Specs.
Tube 6V6 12W
Plate-to-Cathode Voltage: 319V
Tube Max Design Dissipation = 12 watts 100% Pa
Class A Fixed Bias
70% Cool = 26.3mA, 80% Average = 30.1, 90% Max Safe Dissipation =33.9
Class AB Fixed Bias
50% Cool = 18.8mA, 60% Average = 22.6, 70% Max Safe Dissipation = 26.3
Cathode Bias
85% Cool = 32mA, 95% Average = 35.7, 100 % Max Safe Dissipation = 37.6
Ok so I took the measurements and plugged them into Rob's dissipation calculator
shown below are the values and what it calculated. Appears to almost double the max.
:help: :help: :help: Can you see something I did wrong. I only have 1 Output tube in this amp?
Tube Dissipation Using Cathode Resistor Voltage Drop
You must enter Tube Type and Plate-to-Cathode Voltage above before calculating Tube Dissipation.
Enter Number of Tubes that share a cathode resistor: 1
Enter Voltage Across Cathode Resistor: 32.6V
Enter Cathode Resistor: 470 ohms
Total Cathode Current = 69.4 mA
Total Plate Current = 65.6 mA = Cathode Current minus approximate screen current of 5.5%.
Plate Current per Tube = 65.6mA = Total Plate Current / Number of Tubes.
Plate Dissipation per Tube = 20.9 watts = Plate Current per Tube x Plate Voltage.
Plate Dissipation per Tube % = 174.2% Plate Dissipation per Tube
If your tube is a triode (or dual triode with both triodes used) your plate current is
69.4 milliamps, your plate dissipation is 22.1 watts and 184.2 %.
-
69mA is correct, now, what was the voltage you measured on the PS tap that feeds the PA tube? Take that value * ~ .066 and that's YOUR plate diss.
your schematic shows B+1 350ishVDC * .066 = 23W
My recommendation, move the B+1 tap to the right of the choke, replace the choke with a 470ohm 3W. Re-do the measurements.
did the speaker make noise when you were making the measurements?
-
Speaker is quiet and only makes a crackle when turning the power switch on and off. Nothing is plugged into the input jack. I will have to recheck it but I believe it was around 380
-
You can try bumping up Rk (cathode R) to ~750ohms, that drops current, but will bump up plate volts. you want to land somewhere around 12-14W, then fine tune.
so your datasheet 100% shows B+ ~320, current ~ 35mA, that's the target.
-
Ok. Shooter. Tomorrow I'll do some tweaking. Thanks,
-
Bumped the cathode resistor up to 750 ohms to get the cathode current lower.
Plate Voltage = 327 VDC
Cathode Voltage = 21.3 VDC
Plate to Cathode Voltage = 306 VDC
21.3V/750 = .0284A or 28.4mA
327 x .0284 = 9.28W
I haven't removed the choke yet and replaced with the resistor. That will be next.
Since the amp is not over dissipating, guess what? I just had to hear it for a test. Who wouldn't :laugh: It is very very hum free, and sounds awesome.
So dissipate a little more and soon as my tubes get here go from try going from (both halves of the 12AX7) to ( combo of 1/2 of an 12AY7 and 1/2 of a 5751) as a tone experiment. Loving the sound though already. Thanks to all who helped me. I will be sure to keep you all posted on my progress. :icon_biggrin: :icon_biggrin: :icon_biggrin:
One question are wirewound's suitable for cathode resistors?
-
Absolutely.
-
I haven't removed the choke
you're there without swapping it out, leave it if you like it. was just a suggestion to lower B+
-
Ok. Gotcha. But the 750 ohm was correct wasn't it to get a proper safe dissipation percentage. Before that it was over the max.
-
Plate Voltage = 327 VDC
Cathode Voltage = 21.3 VDC
Plate to Cathode Voltage = 306 VDC
21.3V/750 = .0284A or 28.4mA
327 x .0284 = 9.28W
You should use the plate to cathode voltage for your final calculation.
IOW, 306 x .0284 = 8.69W
-
Ok. Cool. I will do that after I add the other tube socket and tweak the resistor for a little more dissipation since I am on the low side right now.
Another question is it a good idea to check any voltages on the preamp tubes. The schematic shows 1.5 cathode volts with 150 Plate Volts. Is it the ratio of those two that is important. I suppose it takes tweeking the plate resistor 100k to get this voltage? Is that important.
-
Ok. Cool. I will do that after I add the other tube socket and tweak the resistor for a little more dissipation since I am on the low side right now.
Another question is it a good idea to check any voltages on the preamp tubes. The schematic shows 1.5 cathode volts. Do I play with the plate 100k to get this voltage? Is that important.
Check these 2 videos out. They will give you some of the back story behind the answer to your question:
https://vimeo.com/28565075
https://vimeo.com/33244955
-
> idea to check any voltages on the preamp tubes.
Your supply rails are zero V and (say) 300V. The tube works with a *part* of the 300V across it and a part across the plate (or cathode) resistor. So as long as the tube plate-cathode voltage is not zero and not full supply voltage, it "is working".
A specific guide: Fender's 12AX7 with Rk=1.5k and Rp=100k, the plate sits near 70% of supply voltage. I would not be distressed about 60% or 80%, except 12AX7 are so consistent around this point that any other result suggests checking resistors and tube. Russian "similar to 12AX7" may be different. 12AU7 *will* bias-up different.
-
Thanks PRR. I just never hear anything on checking them that is why I asked. :icon_biggrin:
-
Ok so I got the amp done with all three tubes now and I am really loving it. Here's the final schematic. I did change up the tone stack also to the original 5F2A tonestack. The 12AY7 and 5751 worked great together.