Epiphone Valve Junior project

[TubeGeek]

I just picked up an Epiphone Valve Junior amp at my local gtr center for $90(it was the floor demo).  It is 5W and has 12ax7 and el84's in it.  I notice that it does not have enough power supply filtering (it has 120Hz hum) so that will have to be changed.  Just out of curiosity I googled this amp and there are already several websites dedicated to modding this amp! It looks like I have a fun weekend project to do.  

One of those websites mentions a lightbulb mod that acts like an attenuator on the output.  Sounds interesting.

I might buy a couple more of them at this price!!!

[TubeGeek]

Another pic

[TubeGeek]

and another...

[tbeck]

What a cool little project!  Looks like plenty of room for a turret board if you wanted to go that far.

[PRR]

OMG!
It's a Champ for $99!

You just have to fix some cheap and stupid details.

[Geezer]

There's some good suggestions for mods in a thread over @ 18w>>>>
http://www.18watt.com/modules.php?name=Forums&file=viewtopic&t=9053

(You must be a member now to even look at the forums now)

"zaphod_phil" has some very good suggestions thru-out  the thread.
I'm going to be near a GC either Monday or Tuesday......I think I'm going to pick up one of the head units! I have a good # of cabs to run it thru.
~$100 for a cab, chassis & iron is pretty amazing.  :o

Jeff

[Tiny_Daddy]

There's also the Galaxy 25 with push-pull output. I got one one of these in the shop with blown power transformer and output tubes. Installed 7591 tubes (6V6 would have worked also) and added a few cathode bypass caps and Hammond power transformer and the owner was pleased as punch.

[TubeGeek]

When I went to gtr center I was there to purchase some picks and I ended up with this amp.  I must say that I am still amazed that it is only $99.  The cabinet, trannies, chassis, and tubes alone are worth more than that.  The speaker is a Weber ceramic 8".

I am thinking of either modifying the stock circuit to correct its inherent problems or maybe building a complete new turret board and dropping that in.  I don't want to turn this into a money pit project but just a simple, quick project. But now that I am thinking of this...which drop in circuit would be suitable for a marshall like dirty tone?

[YD]

What a cool little project!  Looks like plenty of room for a turret board if you wanted to go that far.

Yeah I picked one up the other day too, now I'm just looking for a design to replace that pcb with.  I need a miniz schem now :)

[TubeGeek]

here is a schematic of the circuit...I hope its ok that I post it here.

I post it to open up a discussion on what can be modified to make it sound better...

[Tiny_Daddy]

If you add a filament transformer then you can add a 6.3V pilot lamp and another gain stage, maybe a pentode input. And add some negative feedback.
The 100K resistors that are shown on the filament may really be 100 ohms.

[TubeGeek]

I plan on modding this amp tomorrow so now I just want to iron out the details...

More power supply filtering is needed. I plan on installing two 100uF 350V spragues in parallel with C8 or C6.

Would altering the value of C1 and C2 coupling caps affect tone?  

What would changing the values of r8 and r9 do?  Will it affect gain by changing the bias on the tube? I know that fenders use 1k5 here.

Lastly...I have read that the filaments nees to be changed to either a dc filament design or a twisted wire above the sockets. This is supposed to help eliminate some hum. I will add the extra power supply filtering and then this if there is still hum.

The on/off switch has a pilot light inside of it.

I'm not sure I want negative feedback...I am looking for a dirty tone. Negative feedback would clean it up somewhat, correct?

The 100k's were a mistake and are 100 ohms in the actual circuit.


Can you tell me more about the pentode, more gain mod you speak of?

[Tiny_Daddy]

The coupling caps are purposly small to reduce low frequencies so the amp does not sound too boomy (also called farty). If you want to increase gain, increase the triode plate resistors from 100k to maybe 220K max. Twisted filaments above the sockets is best. Negative feedback makes the amp sound more solid.
I have used Sovtek 7199 in a couple of amps with the pentode plate tied to the triode grid, then 100K triode cathode resistor. Output is off the the triode cathode, wired as a follower with the plate tied directly to tht same B+ that feeds the pentode. If you add this stage I recommend leaving the 100K resistors on the other triode plates or you will get too much gain. You may need to change the 12AX7 to 12AT7 depending on the sound you want. Players say the pentode gives good touch sensitivity, meaning a larger range of sounds depending on how hard you play and where the gain is set on the guitar. You will need to add another filter section to the power supply for the 7199, maybe 22k and 20uF.

[TubeGeek]

I think I will just remove one or two of the existing filter caps and install 100uF 350V in there.  Hopefully that will be enough filtering.  Is there a formula to calculate how much filtering is needed? I know there has to be but I have forgotten those small details...can someone refresh my memory?

I would also like to add a bright cap into the circuit, maybe a 120pF in parallel with r6.  In fact there is a spot for this cap already on the pcb.  I found this by reading another website about mods on this amp.  Is this cap pretty much the same thing as the 120pF bright cap on a fender? By adding this cap are the high frequencies bypassing r6?

Thanks for the replies thus far.

[BriteLite]

This looks to be a simple financially feasible project for me. This morning I discovered a box of stuff I had in a storage unit that contained several filament transformers. Looks like a good project to use one of them.

TubeGeek please keep us advised on what you do.

Peace

[TubeGeek]

Negative feedback makes the amp sound more solid.

more solid???  Do you mean stability in the circuit?

[Tiny_Daddy]

Adding feedback will make the amp sound different.

[PRR]

> which drop in circuit would be suitable for a marshall like dirty tone?

A Marshall.

You are asking how to make a Honda sedan more like a 1-ton pickup. It's not a drop-in. A single-ended amp is a very different beast than the monster Marshalls.

Epi has a push-pull amp at ~$299 that might be the core of a mini-BassMan/Marshall hack.

> More power supply filtering is needed.

Actually, 20uFd is cheap but not stupid.

The bulk of the hum and buzz appears to come from incredibly stupid (or cleverly commercial) heater routing.

Fender Champs work OK with AC heat: the difference is they don't run the heat right NEXT to the plate lead as the Jr does. You'll need DC heat at least on the 12AX7 to get recording-quality low-buzz; properly routed AC heat should be fine for casual knockabout amp use. I think Hoffman has a Champ layout with proper heater routing: cut the PCB pads away from the traces, run proper wires. Ground the now disconnected heater lines on the PCB, this will help suck-out general buzz pickup (since we know these traces work fine for injecting buzz, the reverse is true).

They seem to be conflicted which end of the chassis to ground to. The obvious quick-fix is an insulated input jack, and let ground happen at the output.

The wire from the pot wiper is very critical: if it picks up any signal from the EL84 plate, the amp will get unstable.

[PRR]

The input impedance is absurdly low. It should suck-out the top end of the guitar pickup. The 68K from grid to ground "should" be 1Meg. However, nobody says the amp is dull. It may be that the speaker has a ringy top end, and "normal" pickup loading would make the result too shrill.

> Would altering the value of C1 and C2 coupling caps affect tone?

Not a lot.

C1 passes down to like 7Hz. If you wanted less bass you could reduce this a LOT, but you would get more buzz pickup at that point. Leave it alone.

C2 passes to ~50Hz. You need to shave the bass in a small amp or it will fart when overloaded. You could try 0.01uF for a thinner tone.

C4 and C3 are typical over-kill, flat to ~9Hz. Everybody uses 10uFd-47uFd here for the Normal channel. Brite channels often use 0.68uFd, that's the other obvious value to try for C4. I expect that changing both these caps to 0.68uFd would give no ballz at all.

C5, output cathode cap, looks too small to me. For hi-fi, that would be 100uFd or more. However, hi-fi would use 10 times the output iron; this undersize cathode cap may be needed to prevent fart-bass. OTOH, maybe it could be useful. Throw 100uFd >16V across the existing cap and try it.

[PRR]

> What would changing the values of r8 and r9 do?  Will it affect gain by changing the bias on the tube? I know that fenders use 1k5 here.

Very very slightly. Not worth hacking a PCB. If you toss the PCB and tagboard it, copy a complete Fender preamp.

The minimum-change for more gain is to replace 68K grid to ground with 1 Meg (no input pad and much less pickup loading), short R6, break R7. That's about all there is; it really should be more than enough. Increasing gain does not make sense until you fix the buzz (unless you like to play in the key of 120Hz).

Instead of a monster main cap to smooth the B+ ripple, it is more elegant to add another stage to the filter. The least-hack would be to break the output transformer lead and insert 100 ohms 2 Watts in series, 20-40uFd 400V to ground.

Increasing R13 10K to 22K or 33K might also reduce preamp hum.

R15 seems to be a last-minute hack. A resistor here is an old custom. Values from 220 to 100K give different stability and overdrive effects. For sure a try-it-and-see experiment.

[Lucid_Alice]

Is it wise to add dc filament heater wiring

Like PRR stated, for a knock about amp (Not for a studio recording amp), do a good old fashioned twisted wire for the heaters and get it off the pc board completely. That should bring the 120hz under control. If you're determined to leave the heater wiring on the pc board, then dc may be the only solution worth considering. But that involves a transformer, rectifier, and filtering. Pretty pricey solution for a $99 amp that only needs 2 feet of heater wire to work right.

[bluesbear]

I have a champ clone I made where I used a sheilded twisted pair cable for the heaters. It's next to silent. One of my bands did some recording last week and you can here the results here:
http://thegreywolf.us/iijuicy/IIJuicy_Recordings.htm
Listen to "Blackbird". That's the amp with chorused guitar. No hum!
Dave

[TubeGeek]

Maybe I wasn't clear....I do want this amp to be useable in the studio as that is why I build amps in the first place.  If I cannot use it the studio, then I don't want the amp.  I am just trying to think of all the possibilities out loud and get opinions.

Now the modding begins...I will post my results soon.

[bluesbear]

"Maybe I wasn't clear....I do want this amp to be useable in the studio"

It was clear to me.. I guess I wasn't clear enough. My point is that a sheilded, twisted pair with the sheilding grounded at one end is extremely quiet. I got the idea from one of the TOT books from Donald O'Connor and it works. To really go all the way, use the wire I mentioned . Replace the 2 - 100 ohm resistors with 2 - 50 ohm resistors connected to the outside lugs of a 100 ohm pot and connect the center lug to an approximately 40 volts supply (I use a voltage divider from the B+). It's kind of a pain but it's about as quiet as it gets.
The attached may help with the 40 volt part.
Dave

[Tiny_Daddy]

It don't matter, just git 'er done!

[TubeGeek]

I just finished modding this amp and it sounds much quieter.  There is far less hum.  I installed a 100uF 350V cap in parallel with the c6.  I installed a bright switch with a 120pF cap in the c3 spot.  I also installed a gain switch.  It gives it just a little bump in gain.  I installed a bridge rectifier with a 1000uF cap across the + and - terminals.  From those terminals I connect to the board.  From that position on the board I wired the filaments of the tubes with a twisted pair.  Pins 4 and 5 of the 12ax7 connect to pin 4 of the el84, pin 9 of the 12ax7 connect to pin 5 of the el84.  I also replaced the 0.022 caps with equivalent mallory 150's.  I replaced the 22uF electrolytics with sprague atoms.  

I measure 5.9 VDC on the filaments.

The amp sounds different than stock.  So far it seems that the fizziness of the stock distortion is gone.  It has a nice distortion still.  Now I will play it some more and see if I need to tweak anymore.

[TubeGeek]

Here is the front of the amp with bright switch and gain included.

[TubeGeek]

Here is the inside of the chassis view of the bright switch and gain switch

[TubeGeek]

I notice that there is still hum when the volume pot is maxed out.  Maybe this is unavoidable??? I could replace the input jack with an insulated type and see if that helps.

I am through for tonight but I will try a couple more things that PRR suggested tomorrow.  

I would still like to experiment with lowering the spl of this thing, It is still loud for 5 Watts!

[TubeGeek]

Just out of curiosity I measured the plate voltage of the el84 and it is 354 VDC.  Shouldn't it be around 300 VDC?

If I need to drop the main transformer voltage down is the best way to install a 1k 5w resistor in series from the HV red wire of the PT? There are two red wires...do is matter which wire I install the dropping resistor to? How do I know which one to install the resistor to?

If I understand this all correctly this extra resistor along with the 100UF cap I added makes up an RC filter stage which is good for filtering.  Does this resisor also drop the main voltage down within the el84's proper range, or is that going to be done with another resistor?

PRR said to use a 100ohm 10W resistor in series with the PT red wire...is this going to drop the main voltage?  I asos read that the proper value is 1k here on another forum.

I am trying to sort all this info out and do what is proper. Please help!!!

[PRR]

> plate voltage of the el84 and it is 354 VDC.  Shouldn't it be around 300 VDC?

Plate-cathode or B+? (Though on EL84, there isn't much difference.)

If you want good long life from classic EL84, keep it down to 300V. I suspect that most current production EL84 are either cheap-junk, or use the same 400V-500V plate-metal as the bigger tubes.

If you want maximum power with shorter but still usable life, such as stage guitar, 350V is a fine value for classic EL84.

If you are selling a cheapo amp, you only care that it doesn't fail RIGHT away. What kinda warranty do you get on a $99-$139 tube amp? Even so, classic amp warranties didn't cover tubes.

> do what is proper

But what IS "proper" for a cheap little amp? Maybe "it does what I want"? Do you want 4 Watts or do you want 1 Watt? Do you even know? Awful hard to translate "still loud for 5 watts!" into a target.

Let's see. If an EL84 isn't red-plating at 350V, it must be running 35mA max. The entire output stage works like 350V/35mA= a 10K resistor. If we insert a 100 ohm resistor, 10K/100= 100:1, B+ drops about 1%, power drops 2%. A C-R-C filter with 100 ohms in the middle will do a lot for buzz, not much for excess volume. A 1K resistor will drop about 10% in voltage, about 20% in power, still not much less loud. You might be going to 5K or 10K to seriously reduce volume. At that point you may want to separate the power stage B+ and the preamp B+, so you can drop the power stage to maybe 150V yet keep 300V on the preamp.

> the fizziness of the stock distortion

Could be crappy caps. The electrodes in a cap want to vibrate, much like an electrostatic speaker. Good caps are tightly wound, little vibration. At this box's price-point, they are probably rolling those caps as quick as possible, no matter they come out loose as a roll of economy toilet-paper.

[PRR]

> the new heater wiring...

Too late now.

My thought was to keep the heater wires WAY away from the audio. Take a foot of wire-pair and run from the heater supply the long way around to the back of the PCB and get onto the heater pins without crossing audio paths.

But yeah: even getting a little above the PCB, twisted-pair, with ground-wrap, is a ton better than the original plan. Even at hunret-bux, they could have done better layout. Makes me wonder if they WANTED it to hum, so it didn't steal sales from the $200 and $300 boxes.

[PRR]

.

[tbeck]

Something to think about, if you're getting 354 volts on the plate of the EL84, you better increase the voltage rating of that big sprague cap you installed to 450v, you're right at the edge of exceeding the voltage rating of that 350v cap, it wont be pretty if you blow it up.

[TubeGeek]

I made a couple of changes this afternoon.  I relocated the filament wiring to the open side of the pcb.  This way they have some space instead of under the board scrunched between the chassis and the board, seperated by standoffs. Better.

I removed the 100uF 350V filter cap and replaced that with a 20uF 500V rated Sprague cap. I also added a 470 5W resistor from the + of the 4k7 1W resistor on the pcb to the + lead on the new 20uF cap.  The - lead of that cap connects to the - lead of one of the other filter cap grounds. Better.

I also experimented with r6 and r7, combinations of lower values and shorts and opens and I decided to leave those both at 1M values with r6 having a switchable 1M in parallel. This allows me to switch in a 1M resisotr with r6 which is 1M, thus having 500k when switched on and the gain is perceived to increase.  Better.

I left in the switchable 120pF bright cap.

Thanks to PRR and the rest of you for helping me understand this. I know I am addicted to this forum! I just learn so much. :)

[TubeGeek]

Final...the front panel.

B= bright
N= normal

HI= 500k resistance
LO= 1M resistance

[bigsbybender]

That's just too cool!   By my calculations you're still well under $200 total, you can't beat that!

J.

[TubeGeek]

I am sitting at $110 total. 8-) 8-) 8-)

I am going to purchase a couple more of these tomorrow. Fun weekend project.

[bigsbybender]

Looks like that 20uF 500V sprague was the most "expensive" of the parts ;D  

What did you use for the HI LOW  and Bright/Normal decals....are they regular stickers? If so they look like they're supposed to be there.. nice work.

J.

[TubeGeek]

The sprague was indeed the most expensive part.

I used black rub on lettering from a craft/hobby store.  I have not decided whether to leave it as is or not.

[PRR]

Maybe I'm being obsessive like Nikki Tesla.

> still hum when the volume pot is maxed out.

Some of that is of course the first stage, which has the most gain to the speaker, so the least hum will be audible when the Volume is high.

But I'm looking at your added cap and resistor switching at the top of the volume pot.

The top of the pot is by far the highest impedance in the whole amp. 1 Meg into a 1 Meg pot shunted by 1 Meg, allowing for V1a plate resistance, the top of the pot is 339K impedance, everything else is under 50K (at least at buzz frequencies) (oh, pot wiper is <188K). So the wires to the top of the pot are most vulnerable to buzz.

I think you have added two wires to this node. They are "only" 4 inches, but now you have almost a foot of buzz-catching wire hanging on a 339K node. Shielding will help and also hurt: a foot of shielded wire on this node will cut treble about enough to notice.

Ponder this half-baked plan.

Change R6 to very-low, a jumper or a 1K resistor. Put a new 1Meg R6 right AT the volume pot. If you think vibration is a problem (break the pot leg), use Shoe-Goo to stick the resistor to the pot body. Put your Brite and Gain switches close to the pot (they are probably OK now) and use the shortest possible wires to the top of the pot.

In the hasty-sketch, top is "now" and all the red wires are begging for buzz: several leads passing semi-near the heater supply. Bottom is "revised" and only the short wire(s) tight to the front panel are buzz-catchers.

[Lucid_Alice]

Your second shot at the heater wires look MUCH better. Did you cut the traces for the heater runs on the board? If not they will still be pushing AC around on the circuit board. Measure the traces, they should be dead open.

[TubeGeek]

Thanks.  

I did cut all the traces and I do measure opens.

I cut the traces where they start and end on each trace.

The amp is super quiet except when at 100% maxed.  I don't hear hum when I am playing only when it is idle.

I may try a few more things to get it quieter but not until after I record the amp.

[radar01]

I think I might buy the head version of this thing, gut it, and build a Schedule 40 inside it.

[radar01]

I whipped this up a while back for a GA-5, I'll try this instead of the PCB in the V-Junior.

[TubeGeek]

Do you have a full schematic of what you plan on doing?

I'd be interested in modifying a couple more of these different from each other.

I want to build another just like the one I now have but install a turret board instead of the pcb and keep the circuit components the same.  Just so I can compare one to the other.

[radar01]

Under "Mods and Projects" on the Hoffman Library of Information. Schedule 40 ....

Original 6V6 GA-5 Schematic:

http://www.hoffmanamps.com/projects/images/GA5.jpg

Schedule 40 schematic:

http://www.hoffmanamps.com/projects/images/Schedule40GA-5.jpg

Layout is there too, i think ....

Alan

[radar01]

Everybody here is out of them.  Good luck, and have fun.  Post the results.  You are gonna love the GA-5 (schedule 40) version.  

[bradleyt]

I noticed that the GA-5 used feedback from the speaker output back to the last half of the 12AX7 cathode by way of a 47K resistor.  Is is possible to do this with the valve jr to help reduce some of the noise this amp makes?

So far I've changed the input jack to an insulated one, changed R2 to 33K, pulled R1, put a 1M across the input jack and jumpered the input ground to the pcb ground.  I didn't notice any noise reduction but I am probably hearing heater hum.  That fix is next on the list.

Thanks for any input.

Brad

[radar01]

I would just stick with the Solid State rectifier myself.  Keeps cost low.  The Gibson GA-5 "Les Paul" Reissue uses SS rectifier.  Look at how they did the reissue GA-5 here: http://www.blueguitar.org/new/schem/electar/modnotes/bridged-t_mods_for_lp_jr_amp.pdf

Compare them all, best of both worlds.  Notice the Gibson GA-5 reissue doesn't show NFB loop, old GA-5 does.

My opinion, what makes the GA-5 better than like the Champ 5f1, is the 220K/2.2K plate load / cathode resistor and a little lower value input resistor.  Gibson used twice the resistance on the NFB resistor, 47K as opposed to Fender 5f1 using 22K.  I guess because the GA-5 has twice the plate load resistance on the preamp tubes.  100K on 5f1, versus 220K on GA-5.  GA-5 is pretty crunchy!