Bravo 12AU7 Mods and Appreciation thread

[shattermyillusion]

Well thats good to know =) I'm really loving the way it sounds right now, this is my first headphone amp! Anyway, just curious this is all new to me, what are the benefits for replacing the caps? Is there a 35V cap that would still fit through the top panel?

[Deleted]

It would become a 2200uF/35V or perhaps 3300uF/35V cap.
it's a physical law called the CU product: size=cap x voltage
You can bend the rules a little (by altering the height or width or using other materials)
Or you would have to make the hole bigger so you can fit it.

24Volts across a 25V capacitor is NOT a good thing they are not designed for this, their lifespan will be shortened and they could start to leak over time.
They like to fit as large capacity caps as possible as there seems to be concesus on the web this is the best (not by me in the case of this particular amp, not speaking in general or in the case of lin. PS)

I use a 2200uF/35V but most grotto members believe bigger is better.
Measurements on this amp have shown no degradation of electrical properties even when a 100uF is fitted.
I (personally) could not detect sonic differences either but others have reported noticeable changes.
Has to do with the brain wired/hearing/interpretation/personal points/believes thing I am afraid.

Realise there are also capacitors present in the power supply itself, be they off small capacity in SMPS.
In this case the bigger the capacity the worse the regulation of the SMPS.. few know this though.. has to do with the 'flywheel principle' on which the feedback/regulation SMPS are based UNLESS you also include coils or resistances in the PS lines to isolate them from the regulation circuit.
Does not apply to linear power supplies (transformers, rectifier + regulator)

[jeffc]

I bought one of these Bravo 12AU7 hybrid headphone amps for $50AUS delivered from eBay a while ago and have modded it as per the instructions here to remove cross talk, dim the blue LED, modify resistor values for optimum 12AU7 supply current/life, replace the 1 uf input coupling electrolytics for Wima MKS film types, replace the IRF630 mosfets with IRL types, me being different I decided to use IRL520s instead of IRL510s or IRL530s, why, because specs were inbetween and I could get them from Futurlec for nicks amongst other parts I was ordering. I also added a low-mass 1/8” socked under the board under the 1/4” socket as all my phones have 1/8” plugs fitting and I don’t like using adapters. Also added a 100-120uf photoflash cap under the big reservoir cap, replaced the on/off switch with links and made a couple of other minor changes. Mosfet heatsinks were also changed to higher mass types that provided a large top surface area for adding extra heat sinking on top cut from an old CPU heatsink. The mosfet heatsinks now run only warm rather than too-hot-to-touch. Tube in use at the moment is a clear-top RCA 12AU7 and with this tube the LM317 heatsinks are just warm, maybe 30C. Another story with other tubes but thats another story.

Sound.. absolutely terrific fed from the 32 ohm out of my LiHM battery-powered Trends UD10 USB DAC.

i337.photobucket.com/albums/n390/jeffc_pics/P8011321.jpg[/IMG]
Bravo all lit up with my Trends UD10, short solid 0.5 mm dia silver wire interconnect cable.

Bravo wings and red Wima coupling caps

1/8” socket mounted under the PCB

Mosfet heatsinks

Rear showing switch links and photoflash cap soldered to power socket pins


For the ridiculously cheap price, this little amp is dynamite, delivering very musical sound, and it looks cute too.

cheers .. jeffc

[Deleted]

Nicely done... that should get the temp down, just like Mick's beautifully crafted 'cooling towers'.
Nice mods.... especially the 3.5mm socket below the 6.3 socket is a great thing IMO.

Replacing the polarised electrolytic 1uF capacitors to Polyester types is already a great improvement but since the 1uF coupling caps do nothing else but possibly harm the audiosignal and make a low-cut filter (6.6 Hz - 3dB) at a higher point then the output caps (2Hz -3dB @ 32 Ohm load) I could advise to take them completely out and just put a wire in it's place.
All the other versions (including the G2) are also input capacitor free.
I figure they just wanted to fill the space the cathode pots on the other PCB's took up.

Also you can make the bias voltage adjustable by replacing the 3k0 resistors by 5k pots.
The only major feature these amps lack compared to all the other versions.
This will solve the problem of the LM317's getting hot with other tubes.

If you use isolation sets on the MOSFET's there is no 24 Volts on the cooling fins anymore.

[PinkFloyd]

Aug 1, 2010 5:41:08 GMT jeffc said:

I bought one of these Bravo 12AU7 hybrid headphone amps for $50AUS delivered from eBay a while ago and have modded it as per the instructions here to remove cross talk, dim the blue LED, modify resistor values for optimum 12AU7 supply current/life, replace the 1 uf input coupling electrolytics for Wima MKS film types, replace the IRF630 mosfets with IRL types, me being different I decided to use IRL520s instead of IRL510s or IRL530s, why, because specs were inbetween and I could get them from Futurlec for nicks amongst other parts I was ordering. I also added a low-mass 1/8” socked under the board under the 1/4” socket as all my phones have 1/8” plugs fitting and I don’t like using adapters. Also added a 100-120uf photoflash cap under the big reservoir cap, replaced the on/off switch with links and made a couple of other minor changes. Mosfet heatsinks were also changed to higher mass types that provided a large top surface area for adding extra heat sinking on top cut from an old CPU heatsink. The mosfet heatsinks now run only warm rather than too-hot-to-touch. Tube in use at the moment is a clear-top RCA 12AU7 and with this tube the LM317 heatsinks are just warm, maybe 30C. Another story with other tubes but thats another story.

Sound.. absolutely terrific fed from the 32 ohm out of my LiHM battery-powered Trends UD10 USB DAC.

i337.photobucket.com/albums/n390/jeffc_pics/P8011321.jpg

[/IMG]
Bravo all lit up with my Trends UD10, short solid 0.5 mm dia silver wire interconnect cable.

Bravo wings and red Wima coupling caps

1/8” socket mounted under the PCB

Mosfet heatsinks

Rear showing switch links and photoflash cap soldered to power socket pins


For the ridiculously cheap price, this little amp is dynamite, delivering very musical sound, and it looks cute too.

cheers .. jeffc[/quote]

Very nice job Jeff

[jeffc]

Cheers Mike and thanks for the tips and reminders Frans,

I was wondering about the input coupling caps as I have output caps in the Trends UD10. Agree ... best coupling cap is no cap if it's redundant.

OK next chance I get, I'll removed these Wimas, insert the 5K trimmer pots I have waiting for installation and ensure those Mosfets are isolated, I think they are but will check again. Then just need to set optimum bias voltages - and I'll have a project completed for a change.

I think part of the reason for the great sound this thing is producing, beyond the Frans mods of course, is due the Mosfets not running so blumin' hot, and with the RCA cleartop 12AU7 tube in use, how cool the LM317s also run. I have to leave my finger on each for 10 sec or so before I get any sense of a warm sensation.

cheers..jeffc

[Deleted]

If the LM317's do not get hot they have almost no voltage across them, meaning the bias voltage is probably 12V or lower, and the LM317 is NOT working properly in this case (at least not as a current source).
In it's proper operating point it should get warm or even hot.

[cure]

hello

i just purchased the bravo v2 tube amplifier (which has yet to arrive.)
being a total noob in the tube world, i wanted to ask you guys one quick question:

it says that the compatible tubes for the bravo v2 are 6N10/ECC82, does this mean these are
the only viable tube replacements for the bravo? will it take any 12AU7? say siemens, GE, mullard, telefunken? i mean, tubes will fit into the bravo v2 as long as it's 12AU7?
i've seen alot of 12AU7 being sold on the net, too many variants. i don't have an idea which one would fit the bravo v2.
there's the 12AU7/6189, 12AU7/ECC82, 5814A, E80CC / 6085 (12AU7 compatible), etc....

and what do you mean by 'matched pairs'?

i was thinking of buying tubes so when the bravo arrives, i'll be able to do some tube rolling and test which one suits me best.

again, sorry for these ignoramus questions.

kind regards,

cure

[Deleted]

the Bravo V2/12AU7 is not very suited for tube rolling, at least not as versatile as the G2 (this one has a switch so you can use 6V and 12V heater types) or the other Bravo types that have a little potmeter per channel so you can set the bias.

This amp can ONLY take 12V tubes with v/i characteristics and mu very close to the 12AU7 (= ECC82)
E80CC has slightly different v/i characteristics but might work O.K., optimal bias setting will be off most likely.
Check the datasheets to see if they have a similar footprint (all connections on the same pins) and they must have 12V split heaters (with the split on pin 9)

I recommend the mods by Oohms (first page of this thread) and replacing the 3k0 resistors in the anode CCS for a 5k pot so you can adjust the anode voltage with different tubes when you want to start rolling more tubes.

matched pairs = usually referring to 2 tubes being selected to have as close as possible characteristics (matched pair) for use in stereo equipment where you want L and R to play equally loud/similar.
Also used for describing within 1 tube that the 2 triodes inside it have similar characteristics (matched).
Sometimes the 2 tube halves are not 'close', or with old tubes even far apart.

Had bought a new JJ 803 once where 1 half was bearly working !

[jeffc]

As I had a couple of 12-0-12 30VA encapsulated transformers lying idle I decided to make a 24V regulated linear PSU using a LM317 regulator and fairly standard circuit, soldered point to point on a little vero board from the same batch that I used to make P-to-P Cricket supplies.

I had a higher current rated (4A) LM338 ready for the job, but decided on using a LM317 initially as I have stacks of them and just in case I stuffed up and fried the IC. Lucky ... because I did just that on first power up . Embassing reason , but all other components seemed to survive.

Anyway, it works now and I crafted a larger heat sink to deal with the heat, way too hot to touch using a small one.

How does it sound supplying the Bravo 12AU7 HA? In a word .... excellent ... crystaline clarity and fleshed out textures and a touch of real high-end class compared to the little SMPS.



Initially, seeing I had a plastic case and an earth wire not knowing what to do with, I connected it to the centre tap of the 12-0-12 secondaries, as in a balanced supply, but as I noted some quite audible transformer hum on power up, I disconnected it and have left it floating. Considerably less hum now and only just audible with an ear very close to the transformer.

Best to leave it this way? Any suggestions welcome, Frans/Alex .

cheers.. jeffc

[Deleted]

Jeff
If the transformer has a metal case, you could connect mains earth to one of it's mounting bolts. If not, don't worry about mains earth as it's in a plastic case and appears to be well insulated. BTW, what value mains fuse are you using
Alex

[jeffc]

Alex,

The transformer case is plastic and TBH there was no fuse, good spot Eagle Eyes . I'm not a big fan of them shoot me now but so I can still hang around here and not be sent to the dunces corner or worse I've just inserted a 20mm 500mA semi-delay fuse RS 418-987 on the active, which was a PITA, but now the world is safe and I'm listening to a fuse.

I'd like to see a thread dedicated to the virtues or otherwise of fuses in various situations.

My house has a ELCB, my powerboard as an overload protection circuit breaker and had I left the earth wire connected to the transformer secondary centre tap, surely that would have tripped the ELCB in the event of any downstream shorts. But then again I'm no expert and naivity allows me to walk dangerously on the wild side it seems.

cheers.. Jeff

[Deleted]

Jeff
Electronic gear can suffer terminal damage, instead of readily repairable damage ,if you rely on overload circuit breakers designed to be forgiving of short term very heavy current drains of >10A. e.g. heaters,air conditioners etc..
Alex

P.S.
I am glad that you didn't connect mains earth to the centre tap of the transformer primary instead of the secondary.

[jeffc]

Tinkered with this again on the weekend.

First I checked the tube bias voltage with the beaut RCA clear top 12AU7 I've been enjoying immensely. 9.3 V and 9.4 V each channel so not surprising the LM317 heat sinks were running as cool as a cucumber. Whilst inserting 5K trimmer resistors to fix this, I also shorted across the very close solder pads of the 1uF film input coupling caps. Adjusted the trimmers for 13.5 V, LM317 heat sinks now toasting , checked V again and sat back for a listen. Sound quality a massive step up which I wouldn't have thought possible, absolutely sublime sound quality now for such a cheap amp, completely engrossing and revealing.

But now the rub, without any source input, the volume pot makes a dull rasping sound when turned. Figuring this might be DC, I checked the source output, zilch as expected due to output coupling caps in my Trends UD10 USB DAC. Checked from vol pot wiper to ground and 8 mV DC on one channel, 28 mV on the other for which the rasping sound with no input signal is more prominent.

So maybe those 1 uf coupling caps are important for stopping this small DC that appears to emenate from the amp from affecting the pot? Anyone help here?

And the next question is, will that DC flowing to the vol pot limit its life severely ? I'm reluctant to change a thing though, as in its current state, this little amp sounds quite magical.

cheers.. jeffc

[Deleted]

In a triode there are always electrons making their way from the cathode into the grid.
As small as the grid's wires are they are still present and catch/intercept some electrons that should have made it's way to the anode.
That's why they invented the pentode I would say.
Anyway.. IMO it won't harm the vol pot and is only irritating when adjusting the volume when there is no sound.
The caps (1uF) do block the DC preventing this from happening.
the biggest improvement in sound is because of the proper anode voltage setting in this case.
you can mount the 1uF caps back (I would not use electrolites but non polar caps) or decide to live with it.
You don't have to worry about currents in the volpot.
They are extremely low so won't shorten it's life.

[jeffc]

Thanks kindly Frans,

DC voltages were checked in a rush last night and checking them again on the source input plug this morning "more carefully", the story is...... -0.4/-0.5 mV DC on min vol setting, -52/-81 mV DC on max vol setting. Values being negative. And yes the low level rasping sound is only heard with no signal and not apparent with music and the usually small changes made in listening volumes, so I can live with that. I'm wondering though whether there could be any consequences of this DC output for the source? My Trends UD10 USB-DAC has output coupling caps so no problem here, but other DACs I have don't.

cheers.. jeffc

[Deleted]

the vol pot is likely not to be any higher then 9 - 10 o clock so the DC voltage, on a high ohmic input when nothing is connected, will be around 5-10 mV.

Most sources have a low Ohmic DC coupled output with a certain output resistance of 100 Ohm or lower.
This will load the vol pot at the input as a virtual short and thus the voltage on the input of the amp will be much lower (probably below 0.1mV) because the DC voltage will fall over the unused portion of the volpot (it's a voltage devider) and the output resistor of the source, leaving no trace of it when the output circuit of the source is reached.
This will not harm anything IMO.

[Garage1217]

Hello! I have some questions on the modifications as I have a V2 inbound....

Specifically...
- Do you have a part number for the log style pot for volume? One that you guys prefer?

- How about a part number for reference for the 5K trim pots? Is there a version that fits better than others? I am looking at these...
www.mouser.com/Search/ProductDetail.aspx?R=3362P-1-502TLFvirtualkey65210000virtualkey652-3362P-1-502TLF

- Concerning capacitors, are you only replacing the 6800uF with a 2200uF 35V or are you also replacing the 1000uf units to another value? Was a little confused when reading... Was looking a this unit for the 2200uF
www.mouser.com/Search/ProductDetail.aspx?R=UFW1V222MHDvirtualkey64700000virtualkey647-UFW1V222MHD

- Which capacitors on the board are you getting rid of and running jumpers instead?

- Found the 530's here....
www.mouser.com/Search/ProductDetail.aspx?R=IRL530PBFvirtualkey61370000virtualkey844-IRL530PBF

[Deleted]

5k pot:

that one will do.
It will always need a bit of ingenuity to fit them.
one of the outer contacts and the middle one needs to be connected.
Which one of the outer ones you connect determines if the bias voltage goes up or down when turning clockwise.

The big one needs to be replaced because of it's voltage.
If you can fit a 4700/35V or 3300/35V depends on the diameter and height.

I also replace the 1000 by 2200/25V to lower the cutoff frequency and because a 105oC will las longer near those hot LM317's

I loose the 2 small caps near the volpot and mount a strap instead


You can also use this MOSFET (Mouser nr):
844-IRL510PBF

I would only replace the volpot if the adjustment of the volume isn't going as you would like it to.

[Garage1217]

Great thanks! Just placed the order for the parts. With any luck, they should arrive just before the amp does!

From what I understand these are pretty decent sounding amps with a little work "bravo and indeed" I am surprised no one has made a kit version with a decent PCB so one could tweak as desired or build as desired.

[Garage1217]

Received my Bravo V2 12AU7 model yesterday. Took it out, fired it up and gave it a good few hour listen going back and forth between my zunes direct out and then through the amp. It was not packaged that well, and took over a month to arrive but not that big of a deal as it all arrived in one piece. It looks nice / not bad build quality for the price.

How does it sound? As it was out of the package, all I could think is what a flaming pile of crap it was. It was fat sounding in the bottom end, highs and sparkle did not exists really... was a little harsh... just crappy. All I could think about was how much cash I just wasted on the pile.

So I took a break and came back. Yep still sounded like a flaming pile.

After reading about all the mods, I pretty much performed most all of them.
- Added in 5K trim pots to adjust bias
- Replaced the power supply cap with a 4700iuF / 35V nichicon
- Replaced the two 1000uF units with 2200uF 105C nichicon units
- Yanked out the two small caps on the input side, right next to the tube and ran jumpers
- Replaced the 630's with IRL530PBF mosfets
- And cut some traces and so forth per the crosstalk mod

Everyone kept complaining how horrid and fragile the stock circuit board is. The pics looks the same as the unit I have, except I found it to be reasonable. I was expecting a board made with a laser jet and home kit in which someone sprayed black spraypaint on *LOL* The lead free solder was the only part that made it a pain since it is much higher temp. Just used a solder sucker and cranked up the heat on my station and it removed with ease. Mods were cake overall.

On another note, the stock capacitors were very odd. They weighed about 50% less than the nichicons I used even though they were about the same size. Felt like they were made of foam they were so light? Strange.

Sounds like a lot but all of it took me around 30 minutes. Took me a bit to figure out where to test for bias but got it and set it at 13.75ish vdc on both channels. After this, let it sit for around 30 minutes per the guide I read and re-checked it, made some minor tweaks and tossed on my headphones. In all honesty, I still expected rubbish as I was so disappointed previously. So I listened, and listened some more... Wow! Is this even the same amp? Well not really as I made so many changes but whoa! It really sounds good! The murky thick bottom end was tight and smooth, vocals were fantastic and spacious, the top end that simply wasn't prior WAS! I went back and forth between the direct out of the zune and back to the amp. I was able to get the volume almost perfectly balanced between switching back and forth and the little amp simply made the sound much more enjoyable, especially vocals, mid-bass and bottom end. I had read with these mods, the frequency response was almost flat so I thought I may just be dreaming this up so I had some help and did some back to back a/b to see if I could pick out the little bravo amp and every single time I could so it was not a placebo effect.

Overall I wanted to thank solderdude. and oohms for the posts and writeup. These mods are fantastic and make this cheap budget amp SING! This little guy is the steal of the century for sure!

[Garage1217]

Picking up a tesla / jj 12au7 ecc82 gold pin and a tesla / jj 12au7 spiral filament ecc802 gold pin tube today to try out. I tried a few of my nos 12au7a silvania units last night and did not like them really. Need more time with them to pick out what exactly I did not like. Also have some other 12au7a's to try. Is there any reason the "a" version would be less compatible with this particular amp? Expert opinions welcome as I am still a noob on tubes. Also do particular tubes in the same family like different bias voltages or is it pretty much standard to set them between 13 and 14vdc?

[Garage1217]

Well overall the ecc82 JJ gold pin is my fav without question with the stock china tube coming in second. I really cannot detect a large difference between tubes by my ear except with the yellow box silvania 12au7a. They just sound flat in the vocals for whatever reason. Tried 3 of them. Did not really like the JJ ecc802 either for whatever reason. Anyways here is a pic of my little budget gem.




The JJ ecc82 bias is very strange. It is the only tube I have to turn the pots all the way to the left just to get the voltage to 16V and cannot lower the voltage anymore than that with that particular tube. Both channels are almost dead equal as well and it sounds great! It may be a fluke so I am going to pick up a second today and see if it is the same.

[suloveu]

Hi guys,

this is my first post, but for these couple of months I've been spectating this thread specifically
I've done most of the mods here, and my bravo v2 has sounded far better than it was out-of-box.

Just now, suddenly I got a problem though,
while I was listening to my music over the bravo, some crackling pops occur and suddenly the volume decreased until 50%, all I can notice is that the red LED has gone... I've also measured the voltage across the LED, it is around 7,7V.
Damn I've lost this amp's amazing power...
I can use some helps from you guys though, any suggestions to fix this?

TIA

[suloveu]

I've replaced the red LED, but no luck

I checked and found that the left 2N3906 transistor's all three pins is shorted... Still waiting for the replacement to arrive next week.
Any help / suggestion guys?