Cheap as Chips
-
- Old Hand
- Posts: 693
- Joined: Sun Nov 09, 2008 2:06 pm
- Location: Manchester
#1 Cheap as Chips
So embarking on a budget hybrid amplifier initially for my daughter, but which hopefully might be the basis of a bi-amp option for my open baffles with a bit more gain and umph than my 41Mp/KT66 main amp which would power the Fostex tweeters.
So have bought the lowest cost Chinese kits from e-bay I could find for a valve pre-amp/buffer and LM3886 amplifier, but am thinking of some simple low cost upgrades.
The 6J1 valves are running I think from a 12V ac supply using a voltage doubler and then some form of I think transistor regulator (but with no zener reference). So the absolute max supply voltage would be 34V minus volt drops across diodes, transistors and some in-line resistors. The load resistor I believe to be 4.7k so I can't think that Ia will be much more than 1mA. So at something like 25V Va, Vg might be around -0.4 to -0.5V, obviously much lower than a CD input but some gain should be possible with a potential output voltage swing of +/- 5V. I am told that the valves are wired in triode mode.
First question if I may - I see quite a few of this type of pre-amplifier and headphone amp running off small voltages, well below the normal operating point of say 10mA, 120V and -2V Vg. How might the valves perform audibly at these small currents and voltages? Would the sound quality improve if I managed to raise the operating point towards the more normal levels for this valve?
The 6J1/EF95/CV4010 comes with a 4.7k grid stopper. I have seen other designs with these valves using no grid stopper so should be able to reduce but unsure of by how much as I don't possess a scope, (once read that for most triodes, 200R should be OK). Also what would be the max Rg? Currently fitted with 470k but should I be able to increase this to 1 or even 2M to minimise loss of the input signal? Maybe these are pointless because the available output voltage swing will be the limiting factor.
First mods will be changing the 4 input/output dc decoupling caps, new CV4010 valves and possibly a small CCS load using the BSP129 as I used on my phono amp.
Pics to follow once received.
So have bought the lowest cost Chinese kits from e-bay I could find for a valve pre-amp/buffer and LM3886 amplifier, but am thinking of some simple low cost upgrades.
The 6J1 valves are running I think from a 12V ac supply using a voltage doubler and then some form of I think transistor regulator (but with no zener reference). So the absolute max supply voltage would be 34V minus volt drops across diodes, transistors and some in-line resistors. The load resistor I believe to be 4.7k so I can't think that Ia will be much more than 1mA. So at something like 25V Va, Vg might be around -0.4 to -0.5V, obviously much lower than a CD input but some gain should be possible with a potential output voltage swing of +/- 5V. I am told that the valves are wired in triode mode.
First question if I may - I see quite a few of this type of pre-amplifier and headphone amp running off small voltages, well below the normal operating point of say 10mA, 120V and -2V Vg. How might the valves perform audibly at these small currents and voltages? Would the sound quality improve if I managed to raise the operating point towards the more normal levels for this valve?
The 6J1/EF95/CV4010 comes with a 4.7k grid stopper. I have seen other designs with these valves using no grid stopper so should be able to reduce but unsure of by how much as I don't possess a scope, (once read that for most triodes, 200R should be OK). Also what would be the max Rg? Currently fitted with 470k but should I be able to increase this to 1 or even 2M to minimise loss of the input signal? Maybe these are pointless because the available output voltage swing will be the limiting factor.
First mods will be changing the 4 input/output dc decoupling caps, new CV4010 valves and possibly a small CCS load using the BSP129 as I used on my phono amp.
Pics to follow once received.
TD-125/RB250/MC25FL & 'Snail' phono, NAS/SBT with CS4398 DAC, 41MP pre & MoFo Power, still messing with OBs.
-
- Old Hand
- Posts: 693
- Joined: Sun Nov 09, 2008 2:06 pm
- Location: Manchester
#2 Re: Cheap as Chips
I think the 6P1 valve preamp has a current regulator in the supply, (or it could be a current limiter), in fact it looks like there's one in both the B+ and B- supply rails. Can someone please enlighten me?
If it is a current regulator, am I right in thinking this would forbid/nullify the use of a CSS Anode load?
I'm thinking it would be preferable using an active load on each valve anode, so wondering if I could/should link out the fitted current regulator and then use two CSS Anode loads to maximise gain and reduce distortion? Is this logical?
I was looking for a small current CSS and stumbled on the suggestion in the attachment. If I could use the J112 (part of the 2N4393 family) I think it has an adjustable range from 0.2 to 10mA. Anyone used this type of FET in a CSS?
Need to find my copy of MJ!
If it is a current regulator, am I right in thinking this would forbid/nullify the use of a CSS Anode load?
I'm thinking it would be preferable using an active load on each valve anode, so wondering if I could/should link out the fitted current regulator and then use two CSS Anode loads to maximise gain and reduce distortion? Is this logical?
I was looking for a small current CSS and stumbled on the suggestion in the attachment. If I could use the J112 (part of the 2N4393 family) I think it has an adjustable range from 0.2 to 10mA. Anyone used this type of FET in a CSS?
Need to find my copy of MJ!
TD-125/RB250/MC25FL & 'Snail' phono, NAS/SBT with CS4398 DAC, 41MP pre & MoFo Power, still messing with OBs.
- Mike H
- Amstrad Tower of Power
- Posts: 20189
- Joined: Sat Oct 04, 2008 5:38 pm
- Location: The Fens
- Contact:
#3 Re: Cheap as Chips
The bottom one is deffo a CCS.
The top one looks like a floating regulator, aka 'electronic smoothing choke'. BUT with a current limiter as the second transistor, according to the formula, limits o/p current to 21 mA (by pulling down base of transistor 1), if the resistor across BE of transistor 2 is 33 Ohms.
The top one looks like a floating regulator, aka 'electronic smoothing choke'. BUT with a current limiter as the second transistor, according to the formula, limits o/p current to 21 mA (by pulling down base of transistor 1), if the resistor across BE of transistor 2 is 33 Ohms.
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
- jack
- Thermionic Monk Status
- Posts: 5504
- Joined: Wed Dec 29, 2010 8:58 pm
- Location: ɐılɐɹʇsnɐ oʇ ƃuıʌoɯ ƃuıɹǝpısuoɔ
- Contact:
#4 Re: Cheap as Chips
The top one is not a regulator, it's a capacitance multiplier with a current limiter set (off the top of my head maths) to about 20mA (Vce(sat)/R or 0.7/33 A) - the capacitance multiplier is a common trick to produce an effective capacitance of C1 * T1's beta as seen at T1's output; also, R1 & C1 form a first-order low-pass filter that also effectively reduces the ripple seen by T1's base - With the values shown, the LPF has an Fc of about 0.05Hz, which is rubbish. Another LPF is on the output with an Fc of 0.5Hz so won't do a huge amount at 50 or 60Hz...
Note that the beta for small-signal transistors can vary wildly between samples - the BC337 (the transistor in your schematic) has an hFE (DC current gain) anywhere between 100 and 630 and this directly effects performance Also, driving the base of the pass device directly from Vin will reduce performance & stress it - normally its driven through a voltage divider so the base of T1 sees about 90% or so of Vin.
See http://sound.westhost.com/project15.htm for a better worked example.
The bottom one is a low-current CCS in its simplest form using an N-channel JFET - better (lower noise, higher bandwidth, better stability etc.) CCSs can be made for not a lot more.
HTH
Note that the beta for small-signal transistors can vary wildly between samples - the BC337 (the transistor in your schematic) has an hFE (DC current gain) anywhere between 100 and 630 and this directly effects performance Also, driving the base of the pass device directly from Vin will reduce performance & stress it - normally its driven through a voltage divider so the base of T1 sees about 90% or so of Vin.
See http://sound.westhost.com/project15.htm for a better worked example.
The bottom one is a low-current CCS in its simplest form using an N-channel JFET - better (lower noise, higher bandwidth, better stability etc.) CCSs can be made for not a lot more.
HTH
Vivitur ingenio, caetera mortis erunt
-
- Old Hand
- Posts: 693
- Joined: Sun Nov 09, 2008 2:06 pm
- Location: Manchester
#5 Re: Cheap as Chips
Thanks MikeH/Jack for your replies.
It sounds like the fitted power supply might not be the best but the good news is that if this is a current limiter, I would be able to add CSSs to each anode load. From the link, doesn't sound like it's worth trying to change what's there. Maybe phase 2 might be a new off-board power supply.
You suggest there are better CCS designs - can you please expand? I know of cascading two JFETs and also using a 1k resistor between gate and cathode, but thought keep it simple for this budget design, (and it is a 2-wire solution that would just replace the current 4k7 load resistor). I thought of the 10M45s but read not really OK to use below a few mA. I don't suppose there would be any difference between the J112 and maybe BSP129/135 that I have seen or used before? I did see a 3-wire CCS on an amp called the Millet Minimax using two 2N5087 and an additional ground connection.
It sounds like the fitted power supply might not be the best but the good news is that if this is a current limiter, I would be able to add CSSs to each anode load. From the link, doesn't sound like it's worth trying to change what's there. Maybe phase 2 might be a new off-board power supply.
You suggest there are better CCS designs - can you please expand? I know of cascading two JFETs and also using a 1k resistor between gate and cathode, but thought keep it simple for this budget design, (and it is a 2-wire solution that would just replace the current 4k7 load resistor). I thought of the 10M45s but read not really OK to use below a few mA. I don't suppose there would be any difference between the J112 and maybe BSP129/135 that I have seen or used before? I did see a 3-wire CCS on an amp called the Millet Minimax using two 2N5087 and an additional ground connection.
TD-125/RB250/MC25FL & 'Snail' phono, NAS/SBT with CS4398 DAC, 41MP pre & MoFo Power, still messing with OBs.
- jack
- Thermionic Monk Status
- Posts: 5504
- Joined: Wed Dec 29, 2010 8:58 pm
- Location: ɐılɐɹʇsnɐ oʇ ƃuıʌoɯ ƃuıɹǝpısuoɔ
- Contact:
#6 Re: Cheap as Chips
The current limit in the PSU make no difference to whether or not you can add a CCS elsewhere in the circuit - you can just do it so long as the total current you drawn from the PSU does not trigger the limit, i.e. stays below 20mA.
Initially, I'd stay simple - use the JFET CCS and see how you get on. If that works for you, consider changing one thing at a time - maybe the capacitance multiplier (it's not great) and then the CCS - keep changes simple and well defined, that way you will find what you like the sound of and what's important to you.
Walt Jung, Morgan Jones and others have written about "better" CCS implementations, but I would start simple and then change later if you are unsure.
Initially, I'd stay simple - use the JFET CCS and see how you get on. If that works for you, consider changing one thing at a time - maybe the capacitance multiplier (it's not great) and then the CCS - keep changes simple and well defined, that way you will find what you like the sound of and what's important to you.
Walt Jung, Morgan Jones and others have written about "better" CCS implementations, but I would start simple and then change later if you are unsure.
Vivitur ingenio, caetera mortis erunt
- Mike H
- Amstrad Tower of Power
- Posts: 20189
- Joined: Sat Oct 04, 2008 5:38 pm
- Location: The Fens
- Contact:
#7 Re: Cheap as Chips
Thanks Jack.
This would be better done then with a Darlington type arrangement for T1 ?
It's very similar tho to my so-called FetReg 'electronic choke', which uses a MOSFET.
This would be better done then with a Darlington type arrangement for T1 ?
It's very similar tho to my so-called FetReg 'electronic choke', which uses a MOSFET.
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
- jack
- Thermionic Monk Status
- Posts: 5504
- Joined: Wed Dec 29, 2010 8:58 pm
- Location: ɐılɐɹʇsnɐ oʇ ƃuıʌoɯ ƃuıɹǝpısuoɔ
- Contact:
#8 Re: Cheap as Chips
The bigger the beta the better, but there are more fundamental issues with that design as mentioned. The Elliot Sound Products link I posted above is a good starting point...
Regarding improved CCSs, Walt Jung did a pair of articles in audioXpress in 2007 and a follow-up in April 2009 with a slightly improved design (not that the difference is audible). The PDFs for these are available on his site:
Jung: Sources 101: Part 1 (2007)
Jung: Sources 101: Part 2 (2007)
Jung: Sources 101: Improved version (2009)
Regarding improved CCSs, Walt Jung did a pair of articles in audioXpress in 2007 and a follow-up in April 2009 with a slightly improved design (not that the difference is audible). The PDFs for these are available on his site:
Jung: Sources 101: Part 1 (2007)
Jung: Sources 101: Part 2 (2007)
Jung: Sources 101: Improved version (2009)
Vivitur ingenio, caetera mortis erunt
-
- Old Hand
- Posts: 693
- Joined: Sun Nov 09, 2008 2:06 pm
- Location: Manchester
#9 Re: Cheap as Chips
Thanks Jack.
What current do you think the cascode DN2540 can regulate down to?
Don't quite yet get what he was describing with the various voltage discussions around the choice of J201/J202, but will wait 'til I get the amp and measure the current operating point before considering further.
What current do you think the cascode DN2540 can regulate down to?
Don't quite yet get what he was describing with the various voltage discussions around the choice of J201/J202, but will wait 'til I get the amp and measure the current operating point before considering further.
TD-125/RB250/MC25FL & 'Snail' phono, NAS/SBT with CS4398 DAC, 41MP pre & MoFo Power, still messing with OBs.
- Paul Barker
- Social Sevices have been notified
- Posts: 8996
- Joined: Mon May 21, 2007 9:42 pm
#10 Re: Cheap as Chips
Not very low, that is its shortcoming.
I read the first link Jack gave and it shows a sink working at 2 mA no chance of that with a dn2540.
Think you need to be 10mA and above. But definitely not down in single figures.
I read the first link Jack gave and it shows a sink working at 2 mA no chance of that with a dn2540.
Think you need to be 10mA and above. But definitely not down in single figures.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
- Mike H
- Amstrad Tower of Power
- Posts: 20189
- Joined: Sat Oct 04, 2008 5:38 pm
- Location: The Fens
- Contact:
#11 Re: Cheap as Chips
Why?Paul Barker wrote:shows a sink working at 2 mA no chance of that with a dn2540.
Think you need to be 10mA and above. But definitely not down in single figures.
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
- Paul Barker
- Social Sevices have been notified
- Posts: 8996
- Joined: Mon May 21, 2007 9:42 pm
#12 Re: Cheap as Chips
I don;t know why but it doesn't like low current.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
- Mike H
- Amstrad Tower of Power
- Posts: 20189
- Joined: Sat Oct 04, 2008 5:38 pm
- Location: The Fens
- Contact:
#13 Re: Cheap as Chips
Ah if it's the TO220 jobbie OK, but then I found this: likkle dinky version:
http://www.ebay.co.uk/itm/391015423197? ... EBIDX%3AIT
http://www.ebay.co.uk/itm/391015423197? ... EBIDX%3AIT
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
- jack
- Thermionic Monk Status
- Posts: 5504
- Joined: Wed Dec 29, 2010 8:58 pm
- Location: ɐılɐɹʇsnɐ oʇ ƃuıʌoɯ ƃuıɹǝpısuoɔ
- Contact:
#14 Re: Cheap as Chips
The TO92 version (-N3) has exactly the same silicon as its larger brothers, just lower maximum power dissipation as per the package it's in; electrically and at a silicon level, it's identical. ..Mike H wrote:Ah if it's the TO220 jobbie OK, but then I found this: likkle dinky version:
http://www.ebay.co.uk/itm/391015423197? ... EBIDX%3AIT
Vivitur ingenio, caetera mortis erunt
-
- Old Hand
- Posts: 693
- Joined: Sun Nov 09, 2008 2:06 pm
- Location: Manchester
#15 Re: Cheap as Chips
I may be a little side tracked but got me thinking about my main amp with a single 10m45 running at 20mA .
Is the dn2540 preferred to the 10m45?
I have a few of the latter. Is the consensus that the improved cascode design is worth implementing and assume the same type of circuit would benefit the 10m45?
What sort of audible benefits might this bring about?
Is the dn2540 preferred to the 10m45?
I have a few of the latter. Is the consensus that the improved cascode design is worth implementing and assume the same type of circuit would benefit the 10m45?
What sort of audible benefits might this bring about?
TD-125/RB250/MC25FL & 'Snail' phono, NAS/SBT with CS4398 DAC, 41MP pre & MoFo Power, still messing with OBs.