Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Millett "Starving Student" hybrid amp
New Posts  All Forums:Forum Nav:

Millett "Starving Student" hybrid amp - Page 423

post #6331 of 6756

Edit: Deleted


Edited by Avro_Arrow - 12/3/12 at 8:14am
post #6332 of 6756
Quote:
Originally Posted by Avro_Arrow View Post

Bravo5.jpg

 

This is my somewhat modified Bravo...

 

The schematic is similar to this except the Bravo uses

an LM317 current source instead of the Q2 arrangement.

 

 

700

 

Edit: Had I know this would start a new page I would have used "Quote"

This is in response to the previous post...

Well Why don't I make this a new thread! As I think Modding the Bravo will be the cheapest method [how much did youpay for the parts to mod it]  [although i liked the the guy before us who built the LED lit Millet Hybrid!] Still for a first Tube, I think I'll mod a Bravo that way if I don't like it It should be an easy sell!] 

So to the Led Lit Guy what was your price

post #6333 of 6756
ooh, a bidding war, I like it! No clue.
post #6334 of 6756
Edit: Deleted

Edited by Avro_Arrow - 12/3/12 at 8:14am
post #6335 of 6756

I'm interested in setting up a ccs (one the plates) for the tubes in this amp, however after looking at the datasheets the 12au7s seem to be much more current hungry than their 12ax7 counterparts (about an order of magnitude higher). Could anybody hazard a guess as to what value ccs would be a good starting point for the 12au7s? Would measuring the voltage drop of the current plate load resistor be a good place to start?

 

I would expect that the 12au7s require more current as they have a lower rp, but I don't know if this extra current remains necessary in the ssmh, so ccs advice'd be nice.

 

Cheers,

Chris

post #6336 of 6756
Given the very low currents and voltages involved IMO its very difficult to extract a useful operatng point from the 12AU7 datasheet. So I would say yes: measuring the voltage drop across the plate resistors is a good starting point smily_headphones1.gif If, that is, you like how your amp sounds at its current operating point. If I recall correctly, a 12AU7 build with the default components should be pushing ~0.7mA per tube (0.35 mA per triode)

Another alternative is to simply choose a nice round number, say 1mA, and go for that, which is what I did with my build.

Good luck!
post #6337 of 6756

Hello everyone, I am wanting to make this amp. I am wanting to make the MSSH to drive my Sennheiser HD595s as they lack power and definition when directly hooked up to my computer. I know these are not the best audiophile cans but they were the best I could afford at the time.
 

Anyway, I have been unable to find the_equalizer's version of this amp and the BOM for it. I was planning on making it with the 12AX7, any advice on that would be hanndy if these are the right tubes to use. Also, what is the best way to go about making the amp, should I use some sort of strip board or just solder everything together. I am a novice when it comes to actually building from the ground up electronics.

 

So, any help with this would be greatly appreciated! Thanks!

post #6338 of 6756
Quote:
Originally Posted by the_equalizer View Post

Given the very low currents and voltages involved IMO its very difficult to extract a useful operatng point from the 12AU7 datasheet. So I would say yes: measuring the voltage drop across the plate resistors is a good starting point smily_headphones1.gif If, that is, you like how your amp sounds at its current operating point. If I recall correctly, a 12AU7 build with the default components should be pushing ~0.7mA per tube (0.35 mA per triode)
Another alternative is to simply choose a nice round number, say 1mA, and go for that, which is what I did with my build.
Good luck!

 

Your memory is very close to what I measured today. For the 12AU7 I measured a 20.8Vdrop across the 33k resistor so about 0.63mA of current for the tube. The 12AX7 as expected measured quite differently with a Vdrop of 7.8V so 0.23mA of current. I'm going to be running AU7s so I'll start with 0.75 or 1mA for the CCS as it's still well within the safe plate dissipation from the datasheet graphs and it can't be a bad thing pushing it a bit further up its operating range. Another advantage to the CCS would have to be lower voltage drop before the plates (my 12AU7 plate only sees 24V, it'd be nice to get that up a bit. I'm gonna sim some bits and get on to making this after sorting out my power supply.

 

Quote:
Originally Posted by kidmuffin View Post

Hello everyone, I am wanting to make this amp. I am wanting to make the MSSH to drive my Sennheiser HD595s as they lack power and definition when directly hooked up to my computer. I know these are not the best audiophile cans but they were the best I could afford at the time.
 

Anyway, I have been unable to find the_equalizer's version of this amp and the BOM for it. I was planning on making it with the 12AX7, any advice on that would be hanndy if these are the right tubes to use. Also, what is the best way to go about making the amp, should I use some sort of strip board or just solder everything together. I am a novice when it comes to actually building from the ground up electronics.

 

So, any help with this would be greatly appreciated! Thanks!

I take it you've found the circuit diagram for this version (if not it's here http://diyforums.org/SSMH/SSMHvariants.php along with a bit of information valid for all the variants) as far as I'm aware I haven't come across a BoM for this version however this amp requires very few components and if you read everything of the diagram then you should have everything. With regards to building it, there is a good stripboard design thread around that couple of people have reported success with. I started with stripboard but ended up with point to point over a ground plane as I feel it's better for this particular amp, due to its noise catching properties. IIRC the HD595s are not overly power hungry, you may not need the extra gain that the AXs provide over the AUs then. Well they use exactly the same build, so you can always try rolling them.

post #6339 of 6756
Quote:
Originally Posted by Goobley View Post

I take it you've found the circuit diagram for this version (if not it's here http://diyforums.org/SSMH/SSMHvariants.php along with a bit of information valid for all the variants) as far as I'm aware I haven't come across a BoM for this version however this amp requires very few components and if you read everything of the diagram then you should have everything. With regards to building it, there is a good stripboard design thread around that couple of people have reported success with. I started with stripboard but ended up with point to point over a ground plane as I feel it's better for this particular amp, due to its noise catching properties. IIRC the HD595s are not overly power hungry, you may not need the extra gain that the AXs provide over the AUs then. Well they use exactly the same build, so you can always try rolling them.

Thanks for the advice. I took a look at the diagram and think I got everything sorted. I am wanting to try out different tube and might just start with the 12AU7s for now and later if I want just buy a pair of 12AX7s or 12AT7s.

 

I took the time and created a BOM for this build, roughly based on the original one created by Millett. If there are any suggestion on changes to make it cheaper or even just better, let me know.

 

12au7_mod_starving_bom.xls

post #6340 of 6756

To everyone who helped me build this amplifier, I have good news for you. I am happy to report success. After half a year of debugging, brainstorming and experimenting, the noise level of my starving student is now low enough to be almost inaudible even trough my most sensitive headphones. For the first time since I powered the amplifier for the first time in June, I am currently listening to music trough it and actually enjoying it. And oh my is it great...

 

I came to the conclusion that this amplifier and my DAC simply hate each others. The MSSH gets very noisy when I place it directly on the DAC, but if I leave it on a shelf half a meter away, it's silent.

 

For the first time in half a year since I began this project, I am listening to music trough the MSSH and actually enjoying it. The noise level is now low enough to be forgotten even with my most sensitive headphones. I concluded the MSSH and the NFB-12 hate each others, and they will only be happy if kept apart. This is why I have the MSSH on another shelf instead of stacking it on top the DAC.

 

Last night I debugged the last noise. A mix of flux and dirt got so deep into a tube socket, it allowed current to arc between pins, which made an audible noise. It was a dynamic noise which came and went, cluing me on a possible arcing. By desoldering the interstage coupling cap, then one side of the tube at a time, I managed to narrow down the problem to the socket itself. I already had swapped the tubes and the noise stayed on the same channel, so it wasn't a tube. I soaked the whole sockets in flux cleaner, blew air into the pin holes, and let it dry over night. Today the noise is history.

 

Now the problem is, I made so many modifications simultaneously during my last round of modifications that I have no way of concluding which modification was responsible for lowering the noise... so I'm just gonna conclude all of them had a positive effect.

 

Debugging session in December 2012. There was dirt in a tube socket, which allowed arcing between pins.

 

Currently I don't have the cathode bypass caps in, nor the resistors before the pot. The volume is still usable, but just barely. Since my DAC has a volume control, I use it instead.

 

I think I learned all that was to learn from this project. My next step will be to divide the PSU in two, raise one rail as high as I can for the tubes and lower the other for the output. I've been studying the SOHA II circuitry, and I'll probably aim for something similar. The transformer already has enough voltage on the secondary to let me build a 60 V supply without voltage multiplier, and has lower taps for a 30 V supply. I mean if I'm gonna build a CCS, I may as well go all out. At this point though, I'll stop referring to this amplifier as a Starving Student and start a new thread just for it.

 

Thank you all for the support, you made this project a lot more pleasurable than if I was left on my own. You make this community worth being part of. beerchug.gif


Edited by KimLaroux - 12/27/12 at 1:37pm
post #6341 of 6756

I'm really glad to hear this Kim, I'm also pleased to report that I am in a very similar position to you. After the measurements that I took allowed me to conclude that my power supply was responsible for the noise, I went all out and built a pair of linear regulated power supplies, one at 24V and one at 48V (but with scope for increase of the latter). I've also entirely rebuilt my amp with new components almost everywhere. I managed to finish in time for Christmas when I was very happy to receive a Beyer DT880. What a superb combination these two make...  

 

Just after dinner today I implemented a ccd, based on the_equalizers/minimax design but using BC560s as they were the only transistors I could really get. Using a sensing resistor of 620R my tubes are set to 1.05mA constant current. I may also attempt 1mA per triode as has been suggested. I will also reiterate what has already been mentioned with regards to the crazy extra gain the ccs provides. After turning the amp on, setting the plates to have max voltage and plugging in my headphones I was amazed to hear that I had more background noise than before. It wasn't until I connected a source and listened that I realised why... Previously with my srh440s if I listened at night I could just hear some noise with the pot at minimum, now for the 440s the pot turned all the way down is quite a comfortable listening volume I struggle to get the 250R Beyers over 20° :O  With a gain that high I think a bit of background noise further up the pot (in a totally inaccesible region unless you want your ears to bleed) is pretty damn good.

 

With regards to the cathode bypass caps I would have to say that even though it's something that increases the gain of the amp it does make the bass much creamier, but I'm gonna add some input resistors to try and make the pot usable again, does anyone know what sort of value I can go up to before the sound starts to significantly suffer? (A friend had some 220k input resistors in a different amp and could never get it to sound right 'till he removed them). 

 

Anyhow, I'm going to revive my thread probably tomorrow with some new pictures schematics and info as to what I've done as well as some questions about raising the B+ a notch :)

 

Really glad to hear about your success, this is certainly a wonderful community

 

Cheers,

Chris

post #6342 of 6756

Hey nice to hear you haven't given up! Can't wait to see pictures and schematics...

 

Also nice to see you're also using the 440. Those are so sensitive it's amazing. I'll keep them forever just for this, as they are perfect to debug noises. At the last meet, I used them to prove to other head-fi'ers that even the Bryston BHA-1 had noise on it's output. biggrin.gif

 

Tho mines have the 940's velour pads, which isolate better. Better isolation means you can hear lower levels signal...

 

As for the resistor before the pot, you have to look at them as an extension to the pot's range. Placing a 50 K resistor before a 50 K pot virtually creates a 100 K resistor where only the first half is usable. Well, not exactly half as they are logarithmic, but you get the idea. It simply extends the resistive strip on which the wiper slides. Using a 250 K resistor before a 50 K pot is like using a 300 K pot limited to only 17% of motion. If I had to add a resistor before a pot, I'd start with a low value. For a 50 K, I'd start with something around 22 K. If that's not enough, I'd try 33...

 

But honestly, I just don't like the idea. If the input signal has better SNR than the amplifier, why would you attenuate it and then amplify it again with more noise? It's just silly. Since this amplifier is inverting, you can just run a wire from the output to the input trough a 10K resistor... simple negative feedback that you can adjust. This would be a cleaner solution with other benefits.

post #6343 of 6756

It's great to see that both of you made good progress with your amplifiers and are actually enjoying them now. Congratulations!  It'll be interesting to see how you carry your project further. I believe it has been suggested before (not by me) but I'll reiterate the suggestion here: since you're going with a higher B+, go for another tube. The 12A_7 have never been really good for hi-fi audio, the 6DJ8 on the other hand... :) A bipolar power supply for a complementary MOSFET output stage would do away with the output cap too.... but, hey! I'm starting to list my wishes here!  :)

 

@KimLaroux: That was quite a noise debugging you performed there to diagnose it was a dirty tube socket! 

 

@Goobley: Good to hear you implemented the CCS in your amp. I just love the punch and transient response the amp got with them, particularly with my Shure SRH840. I think you're liking it too :)

 

@Goobley: About the cathode bypass caps, I think you had also mentioned in a previous post that they were related to the frequency response of the tube stage. It actually has nothing to do with increasing the bass response of the tube stage. What the cathode bypass capacitor does is avoid the gain reduction that results from the local negative feedback effect that using cathode resistor biasing implies. Something quite similar to the negative feedback effect the source resistor works in the MOSFET source follower.  The capacitor detours the audio AC signal away from the cathode resistor, hence the voltage across this resistor does not vary, which in turn means the grid to cathode resistor voltage does not vary; hence no local negative feedback effect to reduce the gain. Also, the dynamic changes in gain in an amplification stage are a major source of distortion so, theoretically, a cathode bypass cap also helps improving the linearity of the stage (if we forget about its phase shifiting effects, that is).

 

According to the above discussion, using a CCS as plate load makes the cathode bypass caps completely redundant, because the CCS is already making the bias constant (constant current through the tube means constant current through the cathode resistor, which in turn means constant voltage across said resistor) and getting the maximum gain out of the tube. I must, however, ask other more experienced builders/designers to chime in here, as, for example, the original Millet Hybrid and the MiniMax have both a CCS plate load and cathode bypass caps.

 

About the input resistors to tame the input down. They're certainly not a technically ideal solution but they're a good practical one so if one really wants to tame the gain of the amp down, well, what else is one to do?

 

Anyway, congratulations again, and enjoy your amps and cans!

 

cheers!

post #6344 of 6756

Just a quick word of advice as I'm about to go to bed - almost 1am here...  Just spent a good 4 hours debugging a CCS issue, just before dinner I set the CCS to 2mA and everything seemed to be ok possibly a slightly better sound but marginal. Came back later and noticed that there was a horrible crunching noise in the background, I though it was the CCS then touched the power cable (external psu) and the noise changed significantly, I could almost get it quiet. I had been thinking that the DC barrel jacks that I had used had tabs that were a bit flimsy inside, and when I checked they were almost broken inside, so the ps input is replaced with a terminal block. This however did not fix either the crunch or the variation in its volume when the bales where moved. strange... the connections were very solid after all that and I had used fresh wire, eventually after checking for dry joints in othe rplaces I came back to the CCS and bumped it down to 1mA, the amp is silent again... 

 

I'm guessing that 2mA is too high for the voltage that we're running the valves at and this must be saturation/overdrive (although overdrive's voltage based, isn't it?).

 

 

Second part of this quick note: Kim's idea about NFB from the output of the amp is pretty much genius, (I know it's an old idea but I'd never have thought of implementing it here). With a 20k resistor from the output of the amp to the RCA/pot input I can use around half of the pot with my high impedance cans. As far as I'm aware this should be well below the 20dB level that people say to avoid. To my ears this sounds better than the input resistors that I've played with.

 

The DC barrel jacks I've been using obviously aren't up to scratch (they also had the disadvantage of being interchangeable 24 and 48V), what'd you guys suggest for carrying +48V/+24V/0V/Gnd/and maybe one or two more future voltages to the amp in one connector (if only one is possible)? Possibly something like a PCI-e 6/8 Pin?

 

The CCS was set up more or less as per the_equalizers instructions and with the minimax page http://www.diyforums.org/MiniMAX/MiniMAXccs.php  

Currently I have kept the plate voltage below 50V because I'm using BC560C as my transistors http://www.datasheetcatalog.org/datasheet/motorola/BC560C.pdf  Their V(ce) is -45V max, now the only place where that could be an issue in the ccs would be in QA2 however in my TINA sims and from my scope V(ce) never really seems to go below -10V, since RA9 is not a 'pull-down' resistor as such (I originally assumed it was), would I be ok taking my B+ significantly over 50V (towards 100)?

 

Cheers, thanks again for all the advice,

Chris

post #6345 of 6756

I think you'd have better chances asking about those modifications in another thread... I have a feeling you and I are getting off topic here. In order to respect the original design and keep this thread dedicated to it, I think it would be more appropriate if we continued our discussions in another thread.

New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Millett "Starving Student" hybrid amp