Bravo 12AU7 Mods and Appreciation thread

[vincef71]

After cutting a few traces of copper, I was able to supply the heater with another power supply. The sound quality improved significantly, but it makes my desk a mess with lots of wires. Can you suggest a schematic that I can add to the initial schematic, so it can separate heater supply but with only one power supply to power it all.

I've read a schematic that uses the Traco-2465, but it seems that none of them available locally in my country.

[Deleted]

You have PM

[nickseattle]

Does anybody have problems with very high level of hissing with the amp? Mine is OK with Sennheiser over the ear headphones, but in ear headphones are practically unusable. The hissing doesn't go away even with the volume all the way down and inputs shorted. Tried many different tubes, the same results.

[nickseattle]

Mystery solved. the noise was coming from trim pots. don't buy those!!! Vishay Trimmer Resistor. You wouldn't believe the amount of noise they were creating!!! I'm back to the original R13 & R14, so far they worked quite well with 12au7, but fail sometimes with 5814 tubes.

I made some mods to the amp - cut the heater circuit completely out of the signal path - with its own PS. I expected hum to go down, it did, but boy - TRANSPARENCY WENT THROUGH THE ROOF. I'm not sure why. The amp is becoming my favorite. I tried two ways: 1) battery - 6v - it worked very well, but the obvious issue is that it will run down. 2) second power supply. Couldn't hear a difference in the amount of noise with both methods, so the number 2 is the way to go. I also have Bottlehead Crack and the SQ with the heater out of the way is quite close to Crack, which is amazing!!! With heater in signal path, Crack wins - no questions.

BTW, I would recommend you to try a tube with the amp - Mazda 12au7 - I was absolutely blown away!!! The noise was very, very low, the sound stage is to die for!!! Check it out - it runs about $25 a tube, but well worth it.
tubemaze.info/mazda-12au7-gray-plates/

So now I'm converting Bravo V2 to preamp, so far so good. I'll post my progress.

[Deleted]

It's not the pots that are at fault or introduce the noise.
The noise is determined by the value of the current (resistance value set point) and quality of the tube and is MUCH less with ECC88 types than ECC82,83 types.
Has to do with the applied voltage and transconductance characteristics.

taking the heater out of the audiochain and thereby making sure the current source is always in it's working range over a wide output voltage range is quite measurable (even visible on a scope with triangular waves) and to most people also 'detectable'.

[nickseattle]

solderdude, it's quite possible that I got a bad bunch of the pots. The noise was not constant, but was coming and going and would be very evident when I touch the amp. I narrowed it down to the pots since when I touch them even slightly, the noise would go through the roof.

BTW, I'm trying use the amp as preamp, but experiencing an issue where the amp makes very loud noise - sounds like DC current, if it's was on, but not playing for some time. Very strange - getting concerned with my power amp because of that.

[Deleted]

It sounds like an oscilation or a dodgy tube or connection somewhere to me.
In my experience with dodgy pots they either make HUGE noises which disappear when slightly readjusted or crackle or go completely open circuit on one or more positions resulting in the amp going off with a loud 'sound'.

Is/was the noise in both channels at the same time or only in one ?

Because of the very high amplification factor and the volpot not behind the amp there is a high chance of clipping the amp when something happens.

[pokee1016]

Hello Everyone, I just found this great thread after receiving my Bravo v2 two weeks ago. I have already started modding so I enjoy reading these posts. Am I the only one who started with the power supply? I can tell you with great certainty that a switching supply does not belong in audio and switching to a decent linear supply (preferably regulated) is simply a no-brainer. I built a simple supply with an oversided basic 30V supply and some caps and resistors and immediately the "haze" you get from switching supplies was gone. The other thing I did that was a HUGE improvement was to bypass the 1uf input coupling capacitors with Kimber .1uf which was an eye opener. I'm pulling from parts that I have, so bear with me. Cutting the traces, as suggested here, to de-couple the filaments was a very good mod as well. I don't know what the designer was thinking when he tied them together. I have Silmic II output caps on order as I am certain this will be a huge upgrade over the stock caps in the unit. There's much that can be done here, but starting with the power supply is critical. I have designed audio circuits (mostly tube) for a while now and I can tell you that improving the power supply trumps any improvements you may want to make elsewhere. I enjoy reading everyone's mods. This little thing has potential. I'm just amazed that it runs the 12au7 at such low current and still manages to sound quite good. At least the tube life is limited only to the filament life span.

[leeta]

I just did crosstalk mod and changed to IRF510. I have itunes with as much flac as poss, valabs nosdac, bravo 12au7, grado s80/g-cush... and its sounding sweet. plan to get pots in there to adjust bias current... be great to hear from others still tweaking this little beauty, and anyone using nosdac too.....?

[szabi]

Hi,
Finally I managed to get a few ECC82 tubes, and decided to build a headphone amp around it.
I have read this and the Bravo/Indeed thread, and would like to ask a few questions.
- Removing the heaters from the audio path, will have a drastical impact on SQ? Worth doing it? If yes, what current should i set for the output transistor with the LM317 CCS for best performace?
(I was thinking of using a 7812, that would need to dissipate around 12V * 0.15A = 1.8W)
- If I leave as it is, what are the chances for the LM317's to get shorted IN-OUT, (I'm always worried about getting knock-off parts from our small shops here) and thus burning the filaments - demaging the tube?
- I've read in the other thread that somebody tried out an LM317 in place of FET's. I made some simulations in Tina, and turned out, that it has a much better frequency response. (solderdude, could you make some tests if possible? I am interested in your opinion.)

Thank you.

[Deleted]

heaters in the path = drastic in measurements mainly.
Not many people can hear it though, judging from the huge amount of listeners who listen to stock amps and love it.

LM317 (class-A stage) above 150mA and below 0.5A (depending on heatsinks)

The usage of 7812 as a voltage source for the heater coming from 24V can be done.
Sufficient heatsinks are needed.
the only 2 downsides (aren't killing) is the extra heat produced and the extra current drawn on top of the class-A output stages.

The problem with using LM317 as 'follower' is the relatively high input current.
I made a 'follower' to test a while back and had done some tests regarding input current (adjust pin) using the LM317.
It is too high to be used with the very high Ohmic anode circuit of the tube.
It loads the tube circuit too much.
Works surprisingly well when driven from a low-Ohmic source .
Using an (FET or MOSFET) opamp in front of it may be a solution so you can still make that circuit (using 2 LM317)
Custom Hifi cables used about the same circuit but decided to not use the LM317 as a current source but used a MOSFET and a zener.
In essence 2 FET output stage.

The LM137 can take quite a beating though when loaded incorrectly as it is protected in several ways.

[szabi]

In your modded schematic, the FET current is reduced to 80mA. That's why I was asking.
I would use the 7812 as a separate PS for the filament.

[Deleted]

In the Miridiy it was lowered to 125mA (to limit heat on the small heatsinks), this was possible because I also increased the output resistance.

For the G2 indeed it is 210mA.
This too has to do with the output resistance and intended loads.

re 7812 ... I figured that part out and had added this in the above post.

Welcome to the forum b.t.w.

[szabi]

Thanks Frans for the fast replys. I'm going to buy some parts, and do some tests before I decide which way to go. (btw my HP is a Sennheiser HD428 / 32Ohm)

[Deleted]

If you set the class-A output stage bias to >275mA you can leave out the output resistor if you want.
Heatsinks must be big enough in that case !
That way the output resistance of the amp will be just a few Ohms.

Most 32 Ohm headphones prefer low Ohmic amplifiers.
Some, however, may sound better on a higher Ohmic amplifier.

I have the HD438 (modded) and a few other HP's.

[szabi]

By leaving the cathode resistor (and cap) out, and replacing the anode resistor with a pot, the anode current increases from 270 to 700uA, and therefore the -0.5dB point extends from 52 to 71kHz.
My assumption is, that this way I also would bring noise in from direct ground connection, and would need some sort of grid biasing?
There would be also possible to demage the valve in case of a failure by going into runaway?

(by the way, with a 32ohm load, I get around 16dB of gain)

I have only built one half, and still need to test the separate heater, and some different FET's. (IRF614 so far)

[Deleted]

Oct 8, 2012 11:20:49 GMT szabi said:

By leaving the cathode resistor (and cap) out, and replacing the anode resistor with a pot, the anode current increases from 270 to 700uA, and therefore the -0.5dB point extends from 52 to 71kHz.
My assumption is, that this way I also would bring noise in from direct ground connection, and would need some sort of grid biasing?
There would be also possible to demage the valve in case of a failure by going into runaway?

(by the way, with a 32ohm load, I get around 16dB of gain)

I have only built one half, and still need to test the separate heater, and some different FET's. (IRF614 so far)

Assuming that we are talking about the same schematic,
a better way to increase anode current would be to decrease the value of the emitter resistor in the constant current source transistor.Assuming that the red LED has a forward voltage of say 1.8V, and the emitter to base voltage of that transistor is 600mV, that leaves around 1.2V across the emitter resistor.With a 3K resistor that should result in an Anode current of about .4mA
Alex

Attachments:

[szabi]

Something strange has happend. The anode resistor now is 2.7k fixed, the cathode has a 2.2k trimmer. With an 1kHz sine (or square) from an ipod, I am unable to set the FET source voltage below 12V, even if I short the cathode to ground (cannot make a simmetrical cut)
The simulator however shows, that it should be able to go down to around 5V. (with music, it goes down to 8V)
I've already changed all the components, checked every connection, and still unable to figure out the cause.

As for the separate heater voltage, how much current flows at startup, when the filament is cold?
A damned 7812 did put out (twice) the whole 24V rail on startup. The filament glowed like a lightbulb. Lukily I switcht it right off, don't know how much demage it made, but the valve still works.
(fortunately I have 3 of them)

By the way, I guess, I also prefer the warmer sound of the IRF630.

[szabi]

One more short question, what are the consequences of slightly underpowering the heater? (137mA instead of 150mA)
Probably I will use another (separate) LM317 with a 9.1Ohm resistor as a constant current source for the filament.
Thanks.

[Deleted]

I doubt that there will be any consequences with such a small reduction in heater current with anything but an aging tube.
Alex

[szabi]

The anode resistor now is 2.7k fixed, the cathode has a 2.2k trimmer. With an 1kHz sine (or square) from an ipod, I am unable to set the FET source voltage below 12V, even if I short the cathode to ground (cannot make a simmetrical cut)
The simulator however shows, that it should be able to go down to around 5V. (with music, it goes down to 8V)
I've already changed all the components, checked every connection, and still unable to figure out the cause.

It turned out to be a dodgy valve.
After the heater overvoltage occured with the 7812, I've replaced the tube. Since than, I was unable to set the bias correctly. (Strarted cutting the top side of the sine even If I grounded the cathode)
Interestingly the labels on this tube are barely visible, almost completely washed out. Probably indicating a fair amount of ware and handling.
I swapped with a third one, and everything is fine now.

If you set the class-A output stage bias to >275mA you can leave out the output resistor if you want.
Heatsinks must be big enough in that case !
That way the output resistance of the amp will be just a few Ohms.

Most 32 Ohm headphones prefer low Ohmic amplifiers.
Some, however, may sound better on a higher Ohmic amplifier.

How exactly can I calculate the output resistence of this amp?
The class-A current is now set to 125mA, and the series output resistor is 47 Ohm. (headphones are 32 Ohms)

[leon]

inside bravo ocean

[Deleted]

The redesigned heatsinking is a big improvement ,ASSUMING that they are at least as efficient as the original individual heatsinks,
(Thermal resistance in °C/W) and are able to further reduce internal temperature, but one of the electrolytic capacitors is FAR too close to the valve for longevity.
I am still dubious about it's long term reliability without the usual problems after a relatively short period of use.
Alex

[tomato]

First off, thanks to everyone who contributed and made all these great mods possible. I did not expect it would be this easy to make this little amp sound quite amazing. I just finished changing the power cap in mine and have already done the crosstalk mod and replaced the IRFs accordingly.

I am already quite satisfied with the sound, but I wondered if there is any easy way to lower the overall gain a bit? On my HE-400s it already seems close to overkill on low volumes.

Also, what about changing the blue led to another color led? Would I need to add any resistors or would it just work as is?

I am also still looking for a decent source on heatsinks. Will the heat still be able to be lowered if I just screw another bigger heatsink into the smaller stock ones with some thermal paste between them?

Sorry for asking a lot of beginner questions on just my first post, but this place seems to be the most knowledgeable by far about this little amp.

[twinster]

Hello, I'm expecting a new Bravo Ocean and was wondering if it would be possible to lower the gain of 30 db on it. I'm planning of using it with low impedance can so lower gain would e better.


Thanks