|
|
Post by tuberoller on Feb 5, 2011 13:41:46 GMT
I finally went for all the Solderdude mods and here is the result:  Went a little overboard on the inductor   Changes: 1. Added cathode bias caps 2. Tube roll 3. Changed input cap to 35v / 3300uF / 130ºC 4. Changed output caps to 25v / 2200uF / 105ºC 5. Changed power brick to regulated one (I had lying around) 6. Heater mod A very big thanks to Frans, who helped me through the final stages of the heater mod!! The result sounds amazing... EDIT: After an hour of listening, I can honestly say the heater mod takes the amp to a whole new level. It adds resolution, clarity and power. The noise floor has gone down as well. Highly recommended mod, but not for the faint of heart! EDIT 2: Also changed current source resistors from 5.7ohm to 6.875ohm (22 and 10 in parallel). The Amp runs a lot cooler now. In fact, I find the stock heatsinks on the LM's and IRF510s are more than enough now. |
|
|
|
Post by ozelui on Feb 25, 2011 15:34:27 GMT
Hello everyone!
I have been lurking around on this post for some time and I have finally decided to join!
First of all wanted to thank everyone contributing to Mod/Tweak the Indeed G2.
I will soon intend to modify my own G2 (still waiting to pickup the Fluke...) and apply lessons learned from members such as "admin" aka Rock grotto,"solderdude", "heliharris" and so on...This board rocks!
My two cents: Maybe is that "it´s good cause I´ve done it" thing but I have found useful (GSM noise pickup) to plug RCA caps/covers (Cardas in this case) on currently unused PRE OUTs of the G2...
Does anyone know of any better/bigger heatsinks which will serve as direct replacement for the ones currently fitted to LM317?
Thanks to everyone once again!
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Mar 9, 2011 21:26:33 GMT
|
|
|
|
Post by tuberoller on Mar 12, 2011 12:59:38 GMT
Nice setup you got there Mick!
Reducing the quiescent current by increasing the resistor values below the bias pots (i.e. the last mod in the Solderdude bible) also helps a lot to reduce heat on the G2.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Mar 12, 2011 13:08:35 GMT
Nice setup you got there Mick! Reducing the quiescent current by increasing the resistor values below the bias pots (i.e. the last mod in the Solderdude bible) also helps a lot to reduce heat on the G2. Thanks tuberoller, yes this has the full Frans monte, plus a lot of extra heat sinking, but still gets hot, this base unit works extremely well in conduction with my Ali base plate, i am very pleased with the results. Mick. |
|
|
|
Post by tuberoller on Mar 12, 2011 13:55:01 GMT
I'm going to have a crack at housing the G2 in a case, so will be decreasing the power of the amp even further. Assuming a minimum load of 63 ohm (replacing output resistors from 31,6 to 63 ohm). 7.5V / 63 ohm = 120 mA. So I can get away with 10 ohm bias current resistors, as 1,25V / 10 ohm is still 125mA, still plenty safe to avoid oscillation. The LM's will be mounted against the side of the case, I hope the reduced power will be enough to get away with just a few holes in the top cover to cool down the mosfets. The tube will go on top of the Hammond case, to do this I ordered these parts from a Chinese Electronics monger for cheaps:    DC socket, switch and RCAs will be replaced with panel mounted ones, airwired to the PCB. |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Mar 13, 2011 10:34:43 GMT
Hi tuberoller,
Can you provide a link to the valve guards please (LH pic) please?
TIA,
Dave.
|
|
|
|
Post by tuberoller on Mar 13, 2011 11:07:13 GMT
|
|
|
|
Post by sinnerg on Apr 6, 2011 23:55:44 GMT
I had to sign up just for the thread.  I'm a new owner of a G2. Had it about a week. It's kinda nice, but that crosstalk thing is what led me here. I just by-the-by spotted it on ebay while looking for something else. Been interested in trying a tube amp for ages and it was cheap enough. Really wish I had the money for real good solid equipment, but unfortunately not right now so the likes of the G2 with some tweaking will have to do. That crosstalk is nuts. It was driving me mad. Kinda muddled up the stereo and just felt a little too much. But... I'm a n00b. Like many I'm interested in electronics and have been dying to get into it and understand it. Talk digital and I'm with you ... talk analogue and we're starting to see brain shutdown. How you guys understand all that baffles me completely.  But ... as a n00b I need guidance. Would it be too forward of me to ask for recommendations on what sort of components to get? Just a guideline really as to what you've used in terms of the non-obvious requirements on the components. For example: - power cap -> general purpose, any brand - cathode bypass cap -> low ESR, try panasonic FM or ELNA something-something I think that would greatly help n00bs like me. I see someone used general purpose ELNA for the cathode bypass and that didn't do so well. Also in terms of the IRL510; while I was looking around I couldn't find the IR item and then there is also the IRL510A. I need to pull the spec sheet and see what difference the suffix adds. Ofcourse, as is typical, I can't find this item here. My supplier has the Vishay item available, but no idea how big the difference will be compare to the IR item. I'm also interested in recommendations for online stores that would carry these and not have massive minimum quantity requirements. I've looked at my suppliers and it is even difficult to find the 3.3uF 50V WIMA cap (they only have a 2.2uF 63V but it is special order), let alone trying to find 130C Samwha caps. Where do you source your good components from? On the G2 note, I set the bias on mine this evening. It's just the 12AU7 valve in there (I'm going to be buying a old used E88CC soon and perhaps the E188CC as well). Took a while to figure out how to do the bias on this as you up the one side and the other side comes down. Eventually decided to go up in small increments, matching both sides as I went until I got to 15V. I have to admit, the crosstalk is just about gone. It was originally there from zero on the volume pot and past the 10 o'clock position the whole thing distorted horribly. Now there's no crosstalk up to that point. From 10 on the dial I start to get parts of a distorted feed coming across again. I'm hoping the mods here will fix that all up. It's a cute little amp, not near the level of what I've seen before, but then again it has a certain charm and a pretty good price too if you can't afford the super good stuff. |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Apr 7, 2011 6:58:39 GMT
Crosstalk can only be completely removed with the separate heater power supply modification.
IRL510A is also good to use.
IMO panasonic 105oC caps are excellent value.
beware of old tubes.
Not all are that good anymore, eventhough they are sold as 'good'
I have a 3300uF/35V 105 cap lying around.
Still have 1 (the last) TSR2465 in stock.
|
|
|
|
Post by PinkFloyd on Apr 7, 2011 11:19:47 GMT
I had to sign up just for the thread. I'm a new owner of a G2. Had it about a week. It's kinda nice, but that crosstalk thing is what led me here. I just by-the-by spotted it on ebay while looking for something else. Been interested in trying a tube amp for ages and it was cheap enough. Really wish I had the money for real good solid equipment, but unfortunately not right now so the likes of the G2 with some tweaking will have to do. That crosstalk is nuts. It was driving me mad. Kinda muddled up the stereo and just felt a little too much. But... I'm a n00b. Like many I'm interested in electronics and have been dying to get into it and understand it. Talk digital and I'm with you ... talk analogue and we're starting to see brain shutdown. How you guys understand all that baffles me completely. But ... as a n00b I need guidance. Would it be too forward of me to ask for recommendations on what sort of components to get? Just a guideline really as to what you've used in terms of the non-obvious requirements on the components. For example: - power cap -> general purpose, any brand - cathode bypass cap -> low ESR, try panasonic FM or ELNA something-something I think that would greatly help n00bs like me. I see someone used general purpose ELNA for the cathode bypass and that didn't do so well. Also in terms of the IRL510; while I was looking around I couldn't find the IR item and then there is also the IRL510A. I need to pull the spec sheet and see what difference the suffix adds. Ofcourse, as is typical, I can't find this item here. My supplier has the Vishay item available, but no idea how big the difference will be compare to the IR item. I'm also interested in recommendations for online stores that would carry these and not have massive minimum quantity requirements. I've looked at my suppliers and it is even difficult to find the 3.3uF 50V WIMA cap (they only have a 2.2uF 63V but it is special order), let alone trying to find 130C Samwha caps. Where do you source your good components from? On the G2 note, I set the bias on mine this evening. It's just the 12AU7 valve in there (I'm going to be buying a old used E88CC soon and perhaps the E188CC as well). Took a while to figure out how to do the bias on this as you up the one side and the other side comes down. Eventually decided to go up in small increments, matching both sides as I went until I got to 15V. I have to admit, the crosstalk is just about gone. It was originally there from zero on the volume pot and past the 10 o'clock position the whole thing distorted horribly. Now there's no crosstalk up to that point. From 10 on the dial I start to get parts of a distorted feed coming across again. I'm hoping the mods here will fix that all up. It's a cute little amp, not near the level of what I've seen before, but then again it has a certain charm and a pretty good price too if you can't afford the super good stuff. I can supply the Samwha VA (130C) caps, the WIMA 3.3uF, Panasonic NHG 3300uF / 35V, Panasonic FR 1800uF / 35V, Panasonic FM 1200uF / 35V and a whole raft of other bits and bobs.... probably cheaper than you can get from RS / Rapid etc. What is it you are looking for? Mike. |
|
|
|
Post by PinkFloyd on Apr 7, 2011 20:21:23 GMT
|
|
|
|
Post by sinnerg on Apr 7, 2011 23:22:14 GMT
Yeah, I'll take that.  I found a couple of Farnell affiliates/associates here. I'll give them a call tomorrow to check on their excessive markups. I'll send you a PM in a moment. |
|
|
|
Post by bjarne on Apr 8, 2011 14:04:16 GMT
Hello everybody! First I have to thank everybody making this thread that great! Because of this I bought a G2 a couple of weeks ago instead of a Bravo, which I wanted to buy before. Since then I modded it with FC Caps and a log-pot. It sounds great already, especially with some old Telefunken, Siemens and Tesla Tubes I achieved to get really cheap. In the next couple of weeks I'll do the heater-mod. I got all the parts except the TSR (and the balls to do it). Now my burning question: Is it possible or even a good idea to short the output resistors? I've done much reading on output impedance and I figured I would give it a shot because of my low impedance headphone line-up. But of course I don't wanna destroy my beloved G2.  I'm looking forward to your answer insight-guys! And thx once again for the great work on the G2!!! PS: Has anybody tried a CPU-Thermal-compound on the Mosfets? |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Apr 8, 2011 15:22:13 GMT
It's not a good idea to short the output resistors.
reason: it is a single ended class-A amp and the minumum load resistance it may 'see' is about 50 Ohms.
Since HP's are generally between 32 and 300 Ohms (with some exceptions to 16 or 1000 Ohms) the 30 Ohm output resistor safeguards this.
IF you make the output resistor lower you increase the chance of single sided clipping and because of this DC on the headphone when driven at full power !.
Should you only be using 32 Ohm and higher HP's you could decrease the value to 10 Ohms.
Personally I would increase the value to 56 or 68 Ohm.
You can use thermal compound but won't do much as already some is used on the ones I encountered.
What might be a good addition (the ones I did) I use an insulator set on the IRL's so the heatsinks do not carry +24V anymore decreasing the chance of mishaps.
|
|
|
|
Post by bjarne on Apr 8, 2011 15:53:27 GMT
It's not a good idea to short the output resistors. IF you make the output resistor lower you increase the chance of single sided clipping and because of this DC on the headphone when driven at full power !. Should you only be using 32 Ohm and higher HP's you could decrease the value to 10 Ohms. Personally I would increase the value to 56 or 68 Ohm. Thank you very much Frans! That's exactly what i supposed. What do you mean by "this DC on the headphone"? My HP-Range is from 19 over 32 to 55 Ohm. And all are very efficient with my knob never north of 9 o'clock, even it is a log pot. If I really would short them just to test it, would it be helpful to have a dummy-load on the Jack, so it sees a load without the HP attached? |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Apr 8, 2011 17:08:42 GMT
When you get clipping on 1 side of a sinewave (the negative side) the average voltage will become positive and this will result in' DC like voltages'
If your HP's are low Ohmic and highly effiecient you can even go to 82 or 100 Ohm as an output impedance.
This way you will get more travel on the pot too.
Higher output impedance usually gives improvements in bass and treble response.
The only HP I have that really benefits from a low output impedance of an amp is the Beyerdynamic DT990.
The bass becomes less bloated and the recessed mids come out better.
Other HP's sound better between 50 and 120 Ohm output resistance.
Some can's don't change that much.
The G2 already has a log pot in stock form
The Bravo thingies sometimes have lin vol pots.
If there is an A in the type number it is usually a log pot.
a B or nothing could indicate a lin. vol pot.
|
|
|
|
Post by bjarne on Apr 10, 2011 13:37:18 GMT
When you get clipping on 1 side of a sinewave (the negative side) the average voltage will become positive and this will result in' DC like voltages' If your HP's are low Ohmic and highly effiecient you can even go to 82 or 100 Ohm as an output impedance. This way you will get more travel on the pot too. Now I get it Frans, thank you! Electronic-classes in school are some years back. Yesterday I did an A/B with the G2 and the HP-Plug of my old Yamaha DVD-Player. I suppose the HP-Plug has 120 Ohm, or is there any way to measure from the outside? Anyway, I don't think that a higher Output-Resistor will harm the SQ of my phones. |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Apr 10, 2011 15:10:05 GMT
Most output stages of CD/DVD players are C'Moys actually.
In general, just like C'moy amps the output resistance is very low.
somewhere between 0.5 and 10 Ohm.
Most cans will sound better at higher output resistances of amplifiers but you can ONLY increase the output resistance if there is enough output voltage swing, which is not the case with most inbuilt headphone amps and C'moys such as the Neco amps for instance.
These portable amps usually have better opamps, decoupling and power supply then those that are built in CD/DVD players.
To measure the output resistance of an amplifier/source you will need a resistor of a fixed value ( something between 33 and 100 Ohm ) a CD with a continuous test tone of a few 100 Hz (max 1kHz) and an AC voltmeter or multimeter.
Those multimeters with only 200V and 750V AC settings are not accurate enough at around 1V AC so cannot be used for an accurate reading, you will need a better one with a 2V and /or 20V AC voltage setting on it.
and a calculator....
|
|
|
|
Post by bjarne on Apr 13, 2011 20:50:53 GMT
To measure the output resistance of an amplifier/source you will need a resistor of a fixed value ( something between 33 and 100 Ohm ) a CD with a continuous test tone of a few 100 Hz (max 1kHz) and an AC voltmeter or multimeter. Those multimeters with only 200V and 750V AC settings are not accurate enough at around 1V AC so cannot be used for an accurate reading, you will need a better one with a 2V and /or 20V AC voltage setting on it. and a calculator.... Thanks again! I think I will do some measuring next weekend. After listening quite a while with my Sony CD1700 (32 Ohm) and my AKG K240 MkII (55 Ohm) on the G2, the missing synergy between the Sony and the G2 is freakin' obvious! (Un)Funny thing: with the Sony I can hear some strange HF-Noises on the left channel even with low volumes. I think it's my TFT near to the G2. Does the heater mod help in preventing HF noises? BTW: Does anybody have any experience with the Sony CD3k? It's technically very similar to mine and I thought that maybe anyone know of good synergy with output-impedance. I think I will do the heater-mod in the next few days together with the output current adjustments and some experiments with the output resistors. |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Apr 13, 2011 22:27:23 GMT
Alas no.. If you mount the LM317's on an alu bottom plate the amp does get less sensitive from RF stuff coming from under the amp.
Move it away from the monitor or mount some caps (around 47pF) between the grid and ground.
|
|
|
|
Post by bjarne on Apr 15, 2011 23:15:13 GMT
Move it away from the monitor or mount some caps (around 47pF) between the grid and ground. What do you mean by "between the grid and ground"? Do you mean the distancer-rods by grid?  I will place my Farnell offer in the next days and wanted to know how much Watt do the output resistors need to be? Do I have to order more than 2 of them in order to match them? Would it be a good idea to replace all of those Dale resistors with higher grade ones? I read some bad things about those Dale's. And is it wise to use some silver solder on the components of the G2? And I want to make some custom cable and wanted to use silver solder instead of ordinary one on their connectors. Does it any good on that purpose? |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Apr 16, 2011 7:42:36 GMT
output resistors wattage may be 0.5W to 1W, metal film or carbon as the noise is unimportant at these voltage levels. Don't need to matrch really... even if one channel is 68 while the other is 56 I doubt it can be detected. Another reason being (since there is NO feedback in the amp) channel differences in amplitude between L and R is already MUCH bigger due to the 2 tube halves being different (sometimes even 2dB). Dale are considered to be serious audiophile resistors by many. All resistors (except cheap wirewound) I have ever met are totally flat from DC (0Hz) to at least 1 MHz (1,000 kHz) Yet many claim the sound is different, it not backed up by reason. As you may have noticed I am not a subjectivist so this question is dependent on religion and cannot be answerred by me. If you feel better with changing resistors and caps for certain brands and to use silver solder I highly recommend you do that. At least you will feel you got the maximum SQ out of it and you will always wonder if it could still be better... Also the PCB quality is not that high so replacing parts only once is advised. The 47 pF caps can be placed most easily between the center pin of the volpot and the ground pin of the volpot, those are the pins that are connected together (NOT the hot pin that is connected to the input switch). It is the same as the position I mentioned only at another spot on the PCB. It should prevent HF to enter the tube and being 'detected' by the tube/MOSFET. NOT tested so can't guarantee it will be sufficient. |
|
|
|
Post by bjarne on Apr 16, 2011 19:56:47 GMT
Another reason being (since there is NO feedback in the amp) channel differences in amplitude between L and R is already MUCH bigger due to the 2 tube halves being different (sometimes even 2dB). Dale are considered to be serious audiophile resistors by many. Thank you for your reply! I think I'll let the Dales where they are!  I haven't decided yet wich path in audio I will follow, but your's sound reasonable and due to low budget my whole custom-cable-thing is on ice, so no silver-solder it be! This whole damping factor thing got me bad! And I think there are very reasonable arguments in it's discussion. So I did a bit of research and found this: www.aikenamps.com/GlobalNegativeFeedback.htmWould that be possible and wise on the G2? I think it has lot of pros. And due to the many disposable power on the G2 this wouldn't harm anything, would it? And above it sounded like it would be a good thing to have a feedback in the amp. I'm very anxious for your reply! |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Apr 16, 2011 21:40:57 GMT
I tried that with Ian's Bravo amp and it is possible (applying negative feedback). You would have to insert a resistor between the grid and volpot. Feedback would be dependant on the position of the volume control though Also another resistor + small capacitor which would have to be determined from the grid to the output. IF you were to couple the signal back from before the output resistor and make the total gain of the amp less you would effectively be correcting errors in the amp and output capacitor. in the G2 there is an output relay which should make it possible. on the Bravo this gave problems on start-up (loud 'motorboating type of hum' during startup).  Should you couple back from behind the output resistor this would effectively lower the output resistance of the amp. However, the amplification factor of a single stage amp is not that high and to retain enough amplification AND get a good feedback you would have to use at least 2 tubes. So ... no I wouldn't recommend it for this particular amp, the 'effect' would be too small. For headphones the damping factor is irrelevant as damping does not come from the amplifier but the used diafragm materials and housing/pads not the output resistance (see explanation below for LS). about 30 years ago amps with external feedback inputs (coming from the LS terminals) and damping factors of over 10,000 could be achieved but was abondonned already soon after introduction. For LS a damping factor of 100 is already very good and completely f*#%d up by inductors in the LS themselves and the wiring to it. The damping factor bubble (which was the talk of the town in those days) bursted later. I tell you why it is a myth. the speaker would see a 'short' and thus be damped much better then by an amp with a higher output resistance (tube amps for instance) is the theory. Example: LS 8 Ohm (nominal) amp damping factor 10x so output resistance 0.8 Ohm LS 8 Ohm amp damping factor 100x ( 0.08 Ohm) LS 8 Ohm damping factor 10,000x ( 0.0008 Ohm) Let's assume the internal wiring, low pass filter coil and LS wiring (a few meters) are 0 Ohm (they will be around 0.1 to 0.5 Ohm in practice) What needs to be damped... the speaker being rammed forward by a fast needle pulse needs to be 'braked' so find it's 0 position ASAP. The cone of the woofer has suspension, mass and a tuned air mass behind it which makes it want to resonate at a preferred frequency and it will do so. When a mass with a coil is moved in a magnetic field it generates 'power' (think microphone, generator e.t.c.) This generator has an internal resistance that is equal or bigger then the DC resistance of the coil. Let's assume the speaker generates 8V open voltage (so with no amplifier connected) and has a DC resistance of around 8 Ohm. When tthis coil is shorted with 0 Ohm there would be a current flowing of exactly 1A (8V/8Ohm) in this fictive example which will load the generator (speaker) and thus 'brake' the motion (think car running stationary and switching on the lights and rearwindowheater and notice the engine running slower) so 1A 'breaking power' with an ideal 0 Ohm 0.99999A with the super feedback amp 0.9901 A with DF=100 amp 0.91A with DF=10 amp So a damping factor of 100 (most solid state amps achieve this) will do, especially when one considers the series resistance of the crossover filter and their impedance is already worse then the amp's internal resistance. Ofcourse there will always be experts stating otherwise and websites/forum threads devoted to this holy grail... especially if they can sell you magic... A tube amp with no feedback (G2 and alike) will have more 'tube-like' distortion (read addition of harmonics in a certain amount) then heavily fedback multiple tube amps and thus the G2 will sound more tuby... which is it's fun factor you would be diminishing. Sorry for the long rant.... I hope at least some of it is understood.  |
|
I'm a new owner of a G2. Had it about a week. It's kinda nice, but that crosstalk thing is what led me here.
