Vox AC4 Question(s)

[dwinstonwood]

Hello all,
I'm new here, this is my first post, I'm David.

I'm interested in building my second tube amp. I just finished building an EL34 SET stereo hifi amp - my first ever - that works and sounds great. I had a lot of help on another forum.

I really enjoyed that build, and I'd like to start planning my next project. I really don't want to build a 5F1, or 5E3, etc. I'd like to do something a bit different.

What I'm looking at is the Vox AC4 (V-1-1). However, I'm wondering if it's possible to build it without the tremolo part of the circuit. I've attached the schematic below. I don't know enough about amp circuits to know how much the tremolo circuit interacts with the rest of the amp. I can see that it is connected to the V2 cathode bias for one thing, among others. If I was to build the amp without the tremolo where would I cut out the circuit, and would I need to alter the values of any other components?

Thanks in advance! I really think this would make a cool little 3-tube EL84 amp that's different enough from Champs to have its own sound.

David

[shooter]

dave also 

think this works, but wait for confirmation

[dwinstonwood]

Thank you Dave!

[sluckey]

Shooter got it right but he forgot to cut out the "witch".   

[dwinstonwood]

Thanks sluckey!

I'm going to give this project a shot and start making a bom.

David

[dwinstonwood]

After studying this circuit, I've realized that it's not the best project for a beginner. The rectifier is expensive and uncommon, and I have no idea how to select the best OPT, as well as some of the capacitors.

I'm teaching myself LTSpice (I've been using it less than a week) by copying existing schematics. This schematic has been very frustrating.

So, I'm switching my focus onto the Schedule 40 GA-5. There is a lot more available info on this project, and it still uses an EL84 like the AC4.

First, I'm going to draw the circuit in LTSpice. After hours of trying to understand how to link .SUBCKT text files to an existing component such as the pentode in the misc folder, I discovered that I can open the text file in LTSpice, right-click on the Name line and let the program create the component for me!

I will also continue reading everything on this forum about the Schedule 40, and learn all I can on my own before posting a bunch of questions.

Thanks,
David

[shooter]

I have no idea how to select the best OPT

the tube datasheet holds the key, in this example ~ 5k

[PRR]

> The rectifier is expensive and uncommon, and I have no idea how to select the best OPT

Yeah. If you get it wrong the Earth will turn backward and we will all fly off into space.

BAH! Rectifiers are generic and not sound-shapers. As said there's DATA in datasheets. And this amp was NOT designed in simulation. It was a guy with various parts and a hot soldering iron. (It is really a cross between a good radio and a small PA amp.)

And it is really a Fender Champ with different tubes. Champ transformers WILL work. That US rectifier we like on Champs is all the same to the tone. The EL84 needs a different socket and a *different* cathode bias resistor, just get the right parts. That pentode preamp echos the early Champs but you may as well use the VOX tube and values so "it isn't a Champ".

[PRR]

...think this works, ...

C5 should stay with the "260V" point. It may work without it, but it is Good Practice.

By logic, R7 should be 1.5x-2.0x larger without that 100k dumping into it. I bet it works without this adjustment. The noted DC voltages will be different. But pentode voltages can vary a lot and still work fine.

[dwinstonwood]

Thanks PRR. I'll amend my schematic.

I'm sensing encouragement to stay with the AC4. So, can I just use a "Champesque" 5Y3GT?

[shooter]

So, can I

Yep, can also use SS bridge, diodes.....
you just "match" the transformer to the load's requirements, walla DC power 

[sluckey]

The rectifier is expensive and uncommon

I would use an EZ81. Same pin out, higher current capacity, much cheaper, and it can work on the 6.3vac filament string just like it's little brother EZ80.

[dwinstonwood]

Thanks sluckey.

I will look at the data sheets for the three tubes: EZ81, EF86, and EL84, and see if I can figure out how many amps a PT's 6.3v windings will need to supply power to all three tube's heaters. I think I saw somewhere that the EZ81 should have 1A. For all three tubes, I'm guessing I'll want to choose a PT that's rated at 6.3V @ 3A minimum, maybe 4A?

[tubeswell]

Thanks sluckey.

I will look at the data sheets for the three tubes: EZ81, EF86, and EL84, and see if I can figure out how many amps a PT's 6.3v windings will need to supply power to all three tube's heaters. I think I saw somewhere that the EZ81 should have 1A. For all three tubes, I'm guessing I'll want to choose a PT that's rated at 6.3V @ 3A minimum, maybe 4A?

The EZ81 needs its own separate 6.3VAC winding (i.e. that's two separate 6.3v windings altogether). The EZ81 6V winding wants to be riding on top of the B+ voltage (in order to minimise h-k voltage) - whereas the other 6.3V winding wants to be referenced closer to ground potential (for the same reason).

[dwinstonwood]

Thanks tubeswell. I'll need to read up on that... the part about riding on B+ and heater/cathode voltages went over my head.

Maybe the 5Y3GT on its own separate 5V secondary will be easier for me to comprehend.

I've reached a point with my LTSpice schematic where I want to post what I've done and find out where all of the mistakes are. I could not find a model online for the EF86 that had a fifth pin (for the suppressor grid I guess?) so that component is not going to work in a simulation unless I can figure out how to add it to the .inc file myself.

[PRR]

> EZ81 needs its own separate 6.3VAC winding .... (in order to minimise h-k voltage)

It's rated 500V H-K. It is clearly intended to be used with heater nearer B-, shared with the other tubes.

[tubeswell]

> EZ81 needs its own separate 6.3VAC winding .... (in order to minimise h-k voltage)

It's rated 500V H-K. It is clearly intended to be used with heater nearer B-, shared with the other tubes.

Noted thanks. That aside, would it not be less risky for h-k insulation if it has its own separate filament winding?

[dwinstonwood]

So, this schematic actually runs without error messages in LTSpice.

I have no idea what that means... apparently the app doesn't provide smoke and flame animations.

[tubeswell]

Thanks tubeswell. I'll need to read up on that... the part about riding on B+ and heater/cathode voltages went over my head.

The rectifier tube cathode is  ‘sitting’ at B+ voltage because it is connected directly to the reservoir filter cap. In an indirectly heated rectifier (like an EZ81), the rectifier’s 6.3vac filament is separate from the cathode, but is squeezed right up against the cathode. So an insulation coating compound is applied to the filament that prevents arcing over between the cathode and the filament. This is called the heater-cathode insulation (‘h-k insulation’). The h-k insulation has a maximum rating which if exceeded, will be likely to cause the insulation to fail. (This risk is not a ‘yes/no’ risk, but more of a higher/lower risk. The higher the voltage differential between the filament and the cathode, the greater the risk of h-k insulation breakdown.) h-k failure here will produce a short between the B+ and the filament winding. You can place your trust in the manufactures 500V max h-k insulation rating, and run the EZ81 of the same filament as the other tube heaters, or you can use a separate 6.3VAC source. Up to you.

[kagliostro]

If you want to go traditional, use a vacuum rectifier, but remember that in SE amp you can use Solid State rectifiers without any kind of loss


That is because SE amps idle current Is very close to max power current and B+ did't drop so much like it happens in PPAmps (near NO SAG on SE amps)


Franco

[shooter]

+1
and it's easy on spice, and it's easy to wire and its easy to find a PT....

an SE amp has no "Built in" hum suppression like PP, so the supply should be tight, as little AC riding on the B+ as possible.  A SS FWB with big caps gets you as close as you can get for the least effort.

[PRR]

I've seen many indirect-heat rectifiers serve long life with hundreds of volts of stress.

While there is an internet "fact" that these rectifiers are fragile, I never knew that when working with them. I don't doubt that a few have failed, but how many direct-heat rects fail? (Lots!) Large grain of salt.

The insulation is not crap. It should be good for way more than the usual 100V rating. And we know 12AX7 serves in McIntoshes exposed to 400V swings. I think that 99% of non-rect duties do not need more than a few volts, so they do not promise it, and do not obsess about it in production. But knowing that a specific cathode will be sold with a 500V promise, they may do a full or extra spray/dip, and they may do more arc-testing.

I think there are many other things more worrisome.

[dwinstonwood]

and it's easy on spice

Well, I had to do a lot of Googling, and watch a few videos, but I have a working model.

However, I have no idea which diodes to use, or what resistance and capacitance. I just picked values that I've seen in other circuits.

The green line is input before transformer, the blue line was measured at the capacitor.

[kagliostro]

1N4148 isn't what you want there, it is for other use

1N4007 is commonly used

An upgrade can be 1N5408

Using UF diodes instead of 1N diodes will be a further increase

47uF Is a good value, sometime you see  also 100uF
(higher values can be used but we are talking of a small SE here)

Franco

[shooter]

+1

if you look at the schematic, the cathode lists 8.5vdc and 150ohm R
ohmslaw says 8.5vdc/150 = ~57mA
add a fudge factor to say 70mA
do the ohm thing again using 300vdc (plate) 70mA current to come up with a Resistor
tell you're spice that's the power supply "load"

start with 47uF> 470ohm> 47uF see what you get

[dwinstonwood]

Thank you kagliostro and shooter!

I still need to model a correct PT. I'm not getting the correct voltage at B+ but the schematic is running without errors.

I searched and searched for a 1N4007 model and finally found one. I pasted the .model line into the "standard.dio" file and I guess it's working.

Here is a screenshot with updated components and values:

[shooter]

R1 needs to be in series with the B+ line, NOT to ground, so between the 2 caps +'s

EDIT:
added graphics 
example of a "typical" PS setup from my last build

[dwinstonwood]

Thanks. Here's a 1:4 ratio PT. Blue line measured at C2.

[PRR]

> I still need to model a correct PT.

Why stop there? The PT only reflects what it can suck from the power line. Also limited by the utility generator. How is the boiler fire? Is the waterwheel clogged with fish?

"Normally" it is sufficient to MODEL a B+ with a battery, maybe a few ohms in series.

[shooter]

can you "measure" with spice the AC component on the blue line?  as long as you're playing 

[dwinstonwood]

Thanks PRR. I know I'm not doing this in the most practical way, but the more I put into it the more I learn. Example: I've gained a rudimentary idea of how inductance ratios relate to winding ratios (fwiw).

shooter, not sure if this is what you wanted to see. I seem to have recorded the db levels of ac frequencies:

[shooter]

my twisted logic thinks it's telling me >60hz you're AC is down 100db, in my old life down 120db ticked the box for passes spec 

[dwinstonwood]

It drops off quite a bit right before 60hz:

[shooter]

move on 
your meter and ear will dial it in 

[dwinstonwood]

move on 
your meter and ear will dial it in 

Thanks shooter!

In that case it's time to start ordering parts.

After I get a layout worked up - with some help - maybe someone here can walk me through the process of ordering a custom turret board here at Hoffman.

[shooter]

After I get a layout worked up


 
the .DIY board maker software is pretty easy on a 2 or 3 tube'r, 7 tubes, I'm real busy with OTHER things 

[dwinstonwood]

It looks like I could start with Robinette's "Micro Champ" layout and edit it in DIY Layout Creator, a lot of preliminary work being already done.

[dwinstonwood]

Wow! I thought schematics were challenging to read! Layouts are frying my brain.

[shooter]

Layouts are frying my brain

  If you live in a legal state, it helps 

[dwinstonwood]

  Virginia is getting close, I hear.

OK, some slow progress. Maybe 25% along. I'm sure there are mistakes.

The EF86 has its own "special" pinout that I'm having to learn. Cathode (3) is tied to G3 Suppressor Grid (8).

[sluckey]

The EF86 has its own "special" pinout that I'm having to learn. Cathode (3) is tied to G3 Suppressor Grid (8).

It's acceptable to plagiarize. Here's an EF86 layout that I know works. Maybe it can help with your layout...

     http://sluckeyamps.com/dual_lite/dual_lite.pdf

[dwinstonwood]

Thanks sluckey! That's a lot more elegant than what's on the AC4 schematic. My voltage measurements in LTSpice were all wrong anyway. I'm going to "plagiarizer" it!

I had a difficult time trying to layout the parts, anyway, as you can see:

[PRR]

> My voltage measurements in LTSpice were all wrong anyway.

Between an inexperienced SPICE, and a known-good plan from the Old Days, trust the old info.

I've spent some years trimming SPICE models of triodes for less-wrong results. My pentodes are so far out that I don't even try.

[dwinstonwood]

I can't argue with that PRR! I know my Spice modelling can't be anywhere near correct. Because I really have no idea what I'm doing! But, it's fun trying. Thanks, and it's good to know that the pentode models are suspect. I've tried about three different ones and they all act differently.

So, sluckey, I've put your preamp circuit in my schematic and tomorrow I'll see if I can "translate" it into layout form. I really appreciate the help. I appreciate everyone's help! This is going to be a cool amp.

But, because I can't leave well enough alone I took voltage readings with sluckey's circuit. Everything measured correct except the plate voltage. I changed the 220K resistor to 56K and got the correct voltage.
But, I'm keeping it at 220K for the build. That's the proven real-world value.

[PRR]

> correct except the plate voltage.

Where is the screen voltage?? That is an important part of bias.

As a rough rule: the happy *ratio* of RP to RS is pretty consistently 1:5 for many small pentodes. You seem to be happy with a very different ratio. However you also have a VERY unusual 6.9k RK where 1k-2k is more normal.

[PRR]

.

[sluckey]

So, sluckey, I've put your preamp circuit in my schematic

Your plate resistor and cathode resistor are quite different from my circuit. Gonna have a pretty big impact on operation.

[dwinstonwood]

Now I'm not sure what I posted last night.

I don't think I can take voltage reading in LTSpice, until I can figure out how to set up simulations. I'm doing something completely wrong, and the EF86 model is probably wacky, too. [I get 16.8V on the plate, 270V on the screen, and 6.8V on the cathode.  ]

Anyway, I think this image is correct:

[PRR]

In many SPICES, "1m" is one milli Ohm.

[dwinstonwood]

In many SPICES, "1m" is one milli Ohm.

Holy ____!

I changed the 1M to 1Meg and I have just under 98v on the screen and about 96v on the plate. The cathode is about 2.2v. That got it way closer to sluckey's schematic.

My brain doesn't see the most basic stuff. That would never have occured to me. Thanks PRR.