Millett "Starving Student" hybrid amp | Page 447 | Headphone Reviews and Discussion - Head-Fi.org

vixr · Nov 30, 2013 at 2:35 PM

awesomebob said:
Thats a bit far for me, I live in Italy! But the glue seems to have held, and I can set the distance of the pcb from case by adjusting the nuts, so it seems ok.... But I wish we had radioshack here.....

awesomebob... I think the problem is the length of the bolts you are using. You are able to put a good bit of leverage on the longer bolt, causing them to break the bond with the adhesive. Try using as short a bolt as possible if another one breaks off.

awesomebob · Nov 30, 2013 at 4:13 PM

I agree, its not a solution that I am entirely happy with, but I would like to find a way to avoid having any screws visible on the outside of the case... I have limited skills/tools, resources, and I am reluctant to drill more holes in the case that may be ineffective..... Like you say I think the problem with the screws is the length in proportion to the surface area of the head.... The leverage is too much... I'll figure it out somehow... but any suggestions are welcome!

jshaffe · Nov 30, 2013 at 10:18 PM

I'm not sure if it is because the channels are run using dual triodes but changing tubes in this amp seems to have a lot more effect on the sound than changing tubes in other amps. Has anyone else noticed this?

doors666 · Dec 3, 2013 at 11:54 AM

I am eagerly waiting for the pcb for the 12au7 version.I have the cisco 0.38a supply, headphones are realistic pro30 (32 ohms) and koss porta pro (60 ohms). Chain would be PC->ODAC->B1 preamp->SSMH. I have a few doubts that wanted to clarify while I wait for the pcb to arrive.

1. Is there an error on the 12au7 schematic. why is c1 coming before r13 but c6 is coming from after r13, wont r13 act as a charging resistor for C6. Is this there on pcb also.
2. There is no earth with the psu, should I use 0v for chassis earth also.
3. does this psu have overcurrent protection. Couldnt find a datasheet on the web.
4. do i need to insulate the rcas, headphone out, pot from the chassis.
5. which are good places for carbon film resistors and which are for metal film.
6. can chassis ground point be used as a star ground.
7. Can i safely leave out the pot and control the volume from my B1 preamp.
8. any changes recommended for component values needed for my setup.
9. which caps. can i use:
220uf panasonic FC and FM, Elna RE-3 – which is the best option. Does it help to bypass this with a film cap of say 10-15nf
680uf, 63v – I only got Nichicon PW with this rating, Is it good enough or should I look for something else. Does it help to bypass this with a film cap. I also have 1000uf, 100v Nichicon PJ, can the psu handle this.
470, 63v – only got UCC LXZ. Is it good enough or should I look for something else.
220nf – got wima mkp2 and mks4. I guess mkp2 wins.

00940 · Dec 3, 2013 at 12:33 PM

using this schematic: http://diyforums.org/SSMH/variants/SSMH-12AU7.gif

Quote:

doors666 said:
1. Is there an error on the 12au7 schematic. why is c1 coming before r13 but c6 is coming from after r13, wont r13 act as a charging resistor for C6. Is this there on pcb also.

R13-C6 form a filter, reducing PS noise for the low current stages.

2. There is no earth with the psu, should I use 0v for chassis earth also.
4. do i need to insulate the rcas, headphone out, pot from the chassis.
6. can chassis ground point be used as a star ground.

2. you can connect the chassis to 0V (at only one point preferably).
4-6. not knowing how the pcb is setup, I'll let someone else answers that.

3. does this psu have overcurrent protection. Couldnt find a datasheet on the web.

It would seem so, reading this, "The power supply current limit cycles"

5. which are good places for carbon film resistors and which are for metal film.

r14-r15, r3-r9 are best carbon film, all others can be metal. But you'll probably have no trouble with using all metal film.

7. Can i safely leave out the pot and control the volume from my B1 preamp.

If you do that, you need to wire a 50k resistor from the junction of r16-r14 to ground (same for the other channel)

9. which caps. can i use: 220uf panasonic FC and FM, Elna RE-3 – which is the best option. Does it help to bypass this with a film cap of say 10-15nf
680uf, 63v – I only got Nichicon PW with this rating, Is it good enough or should I look for something else. Does it help to bypass this with a film cap. I also have 1000uf, 100v Nichicon PJ, can the psu handle this.
470, 63v – only got UCC LXZ. Is it good enough or should I look for something else.
220nf – got wima mkp2 and mks4. I guess mkp2 wins.

- 220uf: panasonic FC or FM. In this application, both are good enough (better than the Elna).
- 680uf: the nichicon are fine
- 470uf: the UCC are good (same league as the panasonic)
- 220nf: mkp is usually a bit better.

jddiyer · Dec 10, 2013 at 10:05 AM

I think i've read this thread a dozen times and im feeling lost.
I have ambitiously decided this would be my first diy amp project, it has me drawn in but im yet to finish....
My question is could this drive a very small set of speakers? I was thinking of trying to make a minimalist and very small speaker set to sit with the amp but i really have no idea.

im fine with it only being a headphone amp, i have larger speakers for larger amps anyway. Thanks for your time.

I need to stop second guessing myself, solder everything together and finish this amp!
then move on to the next one, thanks for getting me hooked guys! hahahaha

jshaffe · Dec 10, 2013 at 5:03 PM

jddiyer said:
I think i've read this thread a dozen times and im feeling lost.
I have ambitiously decided this would be my first diy amp project, it has me drawn in but im yet to finish....
My question is could this drive a very small set of speakers? I was thinking of trying to make a minimalist and very small speaker set to sit with the amp but i really have no idea.

im fine with it only being a headphone amp, i have larger speakers for larger amps anyway. Thanks for your time.

I need to stop second guessing myself, solder everything together and finish this amp!
then move on to the next one, thanks for getting me hooked guys! hahahaha

I hooked mine up to a set of logitech speakers and the sound was much less than ideal. It was actually surprisingly loud, but the kind of speakers that are small enough to be powered by this amp are the same kind of speakers that you don't really want to listen to. Plug in a pair of HD600s or HE-400s, you'll be a lot happier with the results.

iso-b · Dec 11, 2013 at 4:46 PM

Hello everyone. This is my first audio project. Lets start with couple photos(shot with iPhone 4), quality is not that good, but I hope that you get idea.

Case is made from Heat threated Magnolia and its coated with clear lacquer. Tubes are 12AU7 from Sylvania. Building case took like 6h and electronic side almost 10h.

Sound is good. There is hum/buzzing sound when volume is turned high enough(without source) and I gets louder when touching pot. Grounding pot shaft would probably solve that. For that other hum/buzz I wont probably do nothing because it wont appear on volume area that I am using while listening. I am not using any grounding plate and desing is point-to-point. Would adding grounding plate solve my problem or is there some other ways to elminate that hum.

jddiyer · Dec 11, 2013 at 10:12 PM

So last night i finished my build (at 2:30am) and got some crackles and very low volume. Im using a old cd player and some simple senn earbuds, but today all im getting is a loud smooth buzz in the right ear and no music at all. I have tried using my phone and other cans but still no luck.

Any suggestions? Tubes are glowing nicely, mofsets are warm without being hot. I do have a DMM but zero idea how to use it :/

Thanks for all the help!

here is a photo of the build, i have a few others on my flickr too. http://www.flickr.com/photos/93935860@N05/11333145253/

soram · Dec 14, 2013 at 1:44 PM

Hi All,
I haven't really listened to music properly in a few years due to disillusionment with a lot of (expensive) equipment and sound quality. One of my loudspeakers gave up the ghost a while back and I've had some temporary replacements that are not great. Rather than spend the money one new ones, a couple of weeks ago, I decided to try some headphones first and see if it was my hearing or the equipment that was the issue! So after some searching I came across this thread and decided to build the amplifier. I also purchased some Sennheiser 650s secondhand.

I bought the components from Farnell, nothing fancy or expensive, a pair of JJ Tesla 12AU7s from Watford valves and wired everything point to point.

I must say I was awestruck at the sound quality! It was spectacular. Deathly quiet and totally engaging. I thought the gain was a little low at first but really I just had it too loud. It is perfectly adequate at around 75% volume.

So I decided to add de_equalizers CCS last week. I got the components as listed and soldered them up, again point to point, turned it on and.......nothing. Well not nothing but very low volume. So I measured up and found the voltage supplying the CCS had dropped to 30V. The addition of the CCS added 7mA across the power supply resistor. I couldn't figure it out as the CCS itself was designed to deliver 1mA only. I eventually realised it was the 10K collector load. I measured 3mA through this. I figured I could just up it so I put in a 220K I had left from the build. This seemed to sort it all out. The voltage came back up to 41.8V and the amp played.

The sound gains were as described. The gain increased, the bass is much tighter and more impressive and I think the upper mid range is much smoother but I need to listen to it a little more to be sure! In any case it is pretty awesome. I would recommend anybody to build it and to add the extra €5 components for the CCS. The BC327s seem to be fine!

That being said I would appreciate any comments on the CCS.

- Is the 220K resistor a bad thing?
- Would it be better to stay at 10K but connect the CCS to the unfiltered power rail instead?
- How did de_equalizer get the measurements with the 10K in place?
- I also put in the trim pots and trimmed the plate voltage to half the supply (20.9V and 41.8V respectively). However, this left me with a cathode voltage of 0.3V and not 0.9V as described in the post. I left it at that rationalising that the plate voltage was more important than the cathode voltage. Is this a bad thing or just a function of my valves?

I guess I will have to get the school books out and make the calculations but I'd prefer just to be listening to it!

Jawmare · Jan 12, 2014 at 5:33 PM

Nevermind, stupid questions.

the_equalizer · Jan 14, 2014 at 10:25 PM

soram said:
Hi All,
I haven't really listened to music properly in a few years due to disillusionment with a lot of (expensive) equipment and sound quality. One of my loudspeakers gave up the ghost a while back and I've had some temporary replacements that are not great. Rather than spend the money one new ones, a couple of weeks ago, I decided to try some headphones first and see if it was my hearing or the equipment that was the issue! So after some searching I came across this thread and decided to build the amplifier. I also purchased some Sennheiser 650s secondhand.

I bought the components from Farnell, nothing fancy or expensive, a pair of JJ Tesla 12AU7s from Watford valves and wired everything point to point.

I must say I was awestruck at the sound quality! It was spectacular. Deathly quiet and totally engaging. I thought the gain was a little low at first but really I just had it too loud. It is perfectly adequate at around 75% volume.

So I decided to add de_equalizers CCS last week. I got the components as listed and soldered them up, again point to point, turned it on and.......nothing. Well not nothing but very low volume. So I measured up and found the voltage supplying the CCS had dropped to 30V. The addition of the CCS added 7mA across the power supply resistor. I couldn't figure it out as the CCS itself was designed to deliver 1mA only. I eventually realised it was the 10K collector load. I measured 3mA through this. I figured I could just up it so I put in a 220K I had left from the build. This seemed to sort it all out. The voltage came back up to 41.8V and the amp played.

The sound gains were as described. The gain increased, the bass is much tighter and more impressive and I think the upper mid range is much smoother but I need to listen to it a little more to be sure! In any case it is pretty awesome. I would recommend anybody to build it and to add the extra €5 components for the CCS. The BC327s seem to be fine!

That being said I would appreciate any comments on the CCS.

- Is the 220K resistor a bad thing?
- Would it be better to stay at 10K but connect the CCS to the unfiltered power rail instead?
- How did de_equalizer get the measurements with the 10K in place?
- I also put in the trim pots and trimmed the plate voltage to half the supply (20.9V and 41.8V respectively). However, this left me with a cathode voltage of 0.3V and not 0.9V as described in the post. I left it at that rationalising that the plate voltage was more important than the cathode voltage. Is this a bad thing or just a function of my valves?

I guess I will have to get the school books out and make the calculations but I'd prefer just to be listening to it!

Sorry for the long time in replying; lately I spend 99% of my time listening to music and only 1% around here

As to the numbers, as mentioned in my original post on the CCS mod, the source of the CCS and values is this section of the MiniMax site. Have a read and see how I arrived at the 1K / 10K resistor values.

A couple of posts below that first post, I did mention that if you set your CCS for higher currents, it would be best to bypass R13, because of the increased voltage drop. In fact I did just that with my amp and it works just fine. Because the CCS loads isolate the triodes from fluctuations in the power supply (great PSRR: Power Supply Ripple Rejection) the RC filter for the triode sections becomes redundant.

Finally, as to the matter of why your CCS are working with that 220 Kohm resistor, I'm at a loss. I would however, advise you to carefully check the circuit layout, wiring and values of the involved resistors in your CSS loads.

Please note that measuring voltages across some parts of the CCS circuit and the triode stage is very tricky because of the high-impedances involved; your meter's input impedance get's clamped in parallel while measuring and can have a very high impact in the measured values. I had to resort to a very old, analog FET input meter and some indirect measurements to reliably 'see' the current values and voltages.

Good luck and keep enjoying your amplifier!

cheers!

soram · Jan 17, 2014 at 6:27 AM

Many thanks for the reply. I wanted to measure all the voltage points again but I too want to spend more time listening! It is a really nice sounding amplifier and spurred me on to make some new power supplies for my LP12 and phono stage! I am again enchanted by the music.....

I am running the same current through the bias but is was the feed to the CCS base drawing too much current. The 220K I think should work but probably won't be as stiff as your design. I will do the calcs over the next few days. I may just bypass the RC filter as you suggest but at the moment the sound is still magic and the bass is totally transparent compared to not having the CCS in place.

My meter is >30 years old! Who knows what input stage it has. I'll check it out before I start measuring away.

Cheers

scorp997 · Jan 31, 2014 at 12:41 AM

Well I finally built a 12au7 version of this amp, I have been watching this thread for a long time.

My brother turned me onto this project and he had built one with the pcb board and is really happy with it. I too purchased a pcb, but once I started looking for a suitable enclosure I changed my mind and went for a point to point build.

What I came up with is shown in my avatar. I can't upload photos yet, since my account is too new. I will post more when I can. The box is from a Kenwood CB-44 Dual Amp Balancer I found at a local PC recycle center and the heat sinks are just clamp-on style CPU heatsinks mounted to the sides of the box. The volume pot was already a 50k and the power switch made a perfect location for the 1/8" headphone jacks. The box is very tight and the way it opens makes it somewhat difficult to keep some of the leads short. I know the heatsinks may work better if turned 90deg, but there is plenty of heat rejection and I feel it looks much better this way. Another really cool thing about this case is that the volume knob has a lit ring around it that will work as a power indicator.

Using all of the optional components, I found the gain was too high when I use my Shure SE315s and an iPod dock outs as a source. I removed the cathode caps and upped the input resistors to 500k and it seems to be working much better. After removing the cathode caps to floor noise is much less, though there is still a little there.

This was a very cool project, thanks to all who made it possible.

the_equalizer · Feb 3, 2014 at 8:13 PM

soram said:
Many thanks for the reply. I wanted to measure all the voltage points again but I too want to spend more time listening! It is a really nice sounding amplifier and spurred me on to make some new power supplies for my LP12 and phono stage! I am again enchanted by the music.....

I am running the same current through the bias but is was the feed to the CCS base drawing too much current. The 220K I think should work but probably won't be as stiff as your design. I will do the calcs over the next few days. I may just bypass the RC filter as you suggest but at the moment the sound is still magic and the bass is totally transparent compared to not having the CCS in place.

My meter is >30 years old! Who knows what input stage it has. I'll check it out before I start measuring away.

Cheers

It's great to read that this little project inspired you and helped keep you enchanted by the music!

If your meter is 30 years old, analog and it does NOT use batteries, then most surely you got mis-leading voltage measurements from it. The problem is the low input impedance of analog, passive meters. We discussed a bit about these problematic measurements back around page 421 in this same thread. You can read more about the meter input impedance problem here and here.

You might want to re-check your measurements to make sure the CCS's are working as intended.

cheers!

Latest Thread Images

vixr

500+ Head-Fier

awesomebob

New Head-Fier

jshaffe

Head-Fier

doors666

New Head-Fier

00940

Headphoneus Supremus

jddiyer

New Head-Fier

jshaffe

Head-Fier

iso-b

New Head-Fier

jddiyer

New Head-Fier

soram

New Head-Fier

Jawmare

New Head-Fier

the_equalizer

1000+ Head-Fier

soram

New Head-Fier

scorp997

New Head-Fier

the_equalizer

1000+ Head-Fier

Users who are viewing this thread