![url]](https://www.diyaudio.com/forums/[url=http://upic.me/show/38637186][IMGDEAD]http://upic.me/i/n3/dsc08288.jpg[/IMGDEAD][/url])
I have performed a huge upgrade of my Fet-10 He, both the unit and the power supply (350VA transformer, 44.000uF Nichicon Gold, etc...).
You can find every detail, including the parts references, on my website:
Threshold Lovers - Threshold - Fet-10 HE 'competition upgrade'
The benefits of this large upgrade is HUGE : with this upgrade, the Fet-10 He is one of the very best preamp that I have ever lessoned to.
I am currently working on a additional upgrade in order to push this preamp even further...
Feel free to ask if you have any question after reading my upgrade. Be aware that additional info has also be given by a DIY audio member on our forum in the " high end preamplifiers" section : www.thresholdlovers.com • View forum - High-end preamplifiers
All the best,
You can find every detail, including the parts references, on my website:
Threshold Lovers - Threshold - Fet-10 HE 'competition upgrade'
The benefits of this large upgrade is HUGE : with this upgrade, the Fet-10 He is one of the very best preamp that I have ever lessoned to.
I am currently working on a additional upgrade in order to push this preamp even further...
Feel free to ask if you have any question after reading my upgrade. Be aware that additional info has also be given by a DIY audio member on our forum in the " high end preamplifiers" section : www.thresholdlovers.com • View forum - High-end preamplifiers
All the best,
Power Supply Upgrade
I've just finished one FET ten/hl project (a Jensen JT-6110K-B based balanced symmetrical output) and I've started planning a power supply upgrade as my next attempt. This is something I've been brooding about for years... since the 1980's.
May I ask what is the current thinking among those who have built upgraded FET ten/hl supplies? I haven't seen new information for a couple of years. From the post dates in different threads I infer GaryB has moved away from favoring a choke input? If so, why? Would people recommend a toroidal transformer or conventional? What diode types are best available for the bridge(s)?
Lastly (for now) I am of the school that ties circuit ground to earth (safety) ground but I see people using resistors or diode bridges to lift their grounds. I understand the point of lifting ground is to minimize ground loops. Can anyone convince me not to ground my ground?
Thanks.
I've just finished one FET ten/hl project (a Jensen JT-6110K-B based balanced symmetrical output) and I've started planning a power supply upgrade as my next attempt. This is something I've been brooding about for years... since the 1980's.
May I ask what is the current thinking among those who have built upgraded FET ten/hl supplies? I haven't seen new information for a couple of years. From the post dates in different threads I infer GaryB has moved away from favoring a choke input? If so, why? Would people recommend a toroidal transformer or conventional? What diode types are best available for the bridge(s)?
Lastly (for now) I am of the school that ties circuit ground to earth (safety) ground but I see people using resistors or diode bridges to lift their grounds. I understand the point of lifting ground is to minimize ground loops. Can anyone convince me not to ground my ground?
Thanks.
Not having to do with my envisioned power supply upgade, I have another FET 10/hl observation:
As expected my JT-6110K-B output resulted in a dead quiet system (not that it was at all bad before). What I was not expecting is (unless I am crazy) a better bass response.
Why could this be? Prior to the transformer the load was 10k. With the JT-6110K-B the load is more like 40k. From the FET 10/h operating manual:
"Your FET ten/h output impedance is 100 Ohms in series with 100 uF. This low impedance assures that components driven by the FET ten/h will not induce errors due to input capacitance or active input error currents. In addition it allows the use of quality cable in fairly extended lengths for connection to the amplifier without the introduction of detrimental signal effects."
If this were true I calculate an output impedance of 100 ohm at 1 kHz and 128 ohm at 20 Hz. Thing is I can't find a 100 uF capacitor in the FET 10/hl. There are four green Roederstein MKT 1813 10 uF 100 v caps that I can see. The chassis is pretty small and if there were a pair of 100 uF output coupling caps in there I should think they would stand out.
If the coupling caps are really 10 uF, as shown on GaryB's schematic earlier in this thread, that would give an output impedance of 101 ohm at 1 kHz and 802 ohm at 20 Hz. I can't think of any other effect that would result in less bass when driving a lower impedance load. Of course it could all be in my head.
Thoughts?
Now, if you'll excuse me, I'm busy relistening to my music collection.
As expected my JT-6110K-B output resulted in a dead quiet system (not that it was at all bad before). What I was not expecting is (unless I am crazy) a better bass response.
Why could this be? Prior to the transformer the load was 10k. With the JT-6110K-B the load is more like 40k. From the FET 10/h operating manual:
"Your FET ten/h output impedance is 100 Ohms in series with 100 uF. This low impedance assures that components driven by the FET ten/h will not induce errors due to input capacitance or active input error currents. In addition it allows the use of quality cable in fairly extended lengths for connection to the amplifier without the introduction of detrimental signal effects."
If this were true I calculate an output impedance of 100 ohm at 1 kHz and 128 ohm at 20 Hz. Thing is I can't find a 100 uF capacitor in the FET 10/hl. There are four green Roederstein MKT 1813 10 uF 100 v caps that I can see. The chassis is pretty small and if there were a pair of 100 uF output coupling caps in there I should think they would stand out.
If the coupling caps are really 10 uF, as shown on GaryB's schematic earlier in this thread, that would give an output impedance of 101 ohm at 1 kHz and 802 ohm at 20 Hz. I can't think of any other effect that would result in less bass when driving a lower impedance load. Of course it could all be in my head.
Thoughts?
Now, if you'll excuse me, I'm busy relistening to my music collection.
Power supplies and output caps / transformers
The previous threads pretty well describe my thoughts on power supplies. The biggest improvement I found was using a really hefty transformer. I think that I ended up with a 300VA transformer. I got really excellent sound with a simple transformer - diode bridge - large cap (20K uf) + bypass caps solution.
Regarding the better bass with an output transformer, that's really interesting. A 4:1 transformer will make a 10K load look like 160K to the preamp. The 10hl has 20uf of output coupling capacitance. At 20hz, the Z of 20uf = 400 ohms. At 100hz, the Z = 80 ohms. For a 10k load, you'll see a 0.27db drop in response due to this from 20hz to 100hz. For a 160K load, you'll see a 0.02db drop due to this effect. Perhaps that's what you're hearing? Or it could be the stronger drive current you're getting due to the transformer. When a transformer drops the voltage by 4x, it also roughly increases current drive by 4x and that might help with the perceived bass. To check this out, you might try increasing the output capacitor size (without the transformer) and seeing how that affects the bass. Some 220uf electrolytic caps would make for an interesting experiment if you used really good quality caps. I've been impressed with some of the recent polymer electrolytic caps.
---Gary
The previous threads pretty well describe my thoughts on power supplies. The biggest improvement I found was using a really hefty transformer. I think that I ended up with a 300VA transformer. I got really excellent sound with a simple transformer - diode bridge - large cap (20K uf) + bypass caps solution.
Regarding the better bass with an output transformer, that's really interesting. A 4:1 transformer will make a 10K load look like 160K to the preamp. The 10hl has 20uf of output coupling capacitance. At 20hz, the Z of 20uf = 400 ohms. At 100hz, the Z = 80 ohms. For a 10k load, you'll see a 0.27db drop in response due to this from 20hz to 100hz. For a 160K load, you'll see a 0.02db drop due to this effect. Perhaps that's what you're hearing? Or it could be the stronger drive current you're getting due to the transformer. When a transformer drops the voltage by 4x, it also roughly increases current drive by 4x and that might help with the perceived bass. To check this out, you might try increasing the output capacitor size (without the transformer) and seeing how that affects the bass. Some 220uf electrolytic caps would make for an interesting experiment if you used really good quality caps. I've been impressed with some of the recent polymer electrolytic caps.
---Gary
I decided to measure the difference in bass rolloff, with and without transformer coupling, by placing a meter (Agilent 344005A) across one speaker while playing test tones. The short answer is that I could measure no difference in rolloff. I am prepared to admit that the effect, if any, is in my head.
Noise, however, measures lower with transformer coupling.
As to the load the FET ten/hl sees with transformer coupling, the JT-6110K-B requires a resistive load of 2200 ohms, so reflected back 16x2200 is 35,200 ohm.
Noise, however, measures lower with transformer coupling.
As to the load the FET ten/hl sees with transformer coupling, the JT-6110K-B requires a resistive load of 2200 ohms, so reflected back 16x2200 is 35,200 ohm.
Ooops...one too many zeros in the Agilent part number: 34405A.
Edit: which is the more recent design of the FET ten/hl, mine with the four green Roederstein MKT 1813 10 uF 100 v caps or the one with two green caps and two yellow? Was there ever a version with 100 uF that the manual gives as the specification?
Edit: which is the more recent design of the FET ten/hl, mine with the four green Roederstein MKT 1813 10 uF 100 v caps or the one with two green caps and two yellow? Was there ever a version with 100 uF that the manual gives as the specification?
Last edited:
Try it without the 2200ohm load and see what you think. I've found that tweaking the loading of a transformer is often helpful for getting the best sound. A set of clip leads and a bunch of resistors let you quickly compare and see what sounds good in your system.
---Gary
I'm reasonably convinced the secondary load value would make a difference in the sound. However Jensen specifies 2200 ohm for flat frequency response, so that is what I used (two 1.1 k resistors). I suppose I could parellel other values to lower the resistance, but I don't want to remove the 1.1 k resistors for fear of over heating them or getting the transformer leads too short. Emphasis was on getting the leads as short as possible, not on provision for disassembly. My understanding is that if the secondary of an input transformer is under damped it rings. Not to mention that the center of the resistor pair is the ground point.
If I had been able to measure a difference in frequency response, I might be more interested to try to change the response. I am grateful for your suggestion even though I probably will not try it.
If I had been able to measure a difference in frequency response, I might be more interested to try to change the response. I am grateful for your suggestion even though I probably will not try it.
I think that the transformer that you're using has a center tap for the output. It would be the position where the orange and red leads are connected together. It would be better to use that as the ground connection instead of the artificial ground you've created with the resistor pair.
---Gary
Some time ago I almost started a thread to ask about this question! I saw three possibilities:
1.) Tie transformer center tap to ground, as you suggest.
2.) Tie resistor pair center tap to ground, which is what I have.
3.) Tie both transformer center tap and resistor pair center tap to ground.
For this application I am more concerned with balance to ground than I am with signal symmetry, though symmetry is nice. I've read that the secondaries of input transformers are not necessarily well balanced. Though I do not know this as a fact.
Why would it be better to use the center tap of the secondary as ground? Possibility 3.) above would not be too hard for me to test, I suppose, if there is a reason for it.
1.) Tie transformer center tap to ground, as you suggest.
2.) Tie resistor pair center tap to ground, which is what I have.
3.) Tie both transformer center tap and resistor pair center tap to ground.
For this application I am more concerned with balance to ground than I am with signal symmetry, though symmetry is nice. I've read that the secondaries of input transformers are not necessarily well balanced. Though I do not know this as a fact.
Why would it be better to use the center tap of the secondary as ground? Possibility 3.) above would not be too hard for me to test, I suppose, if there is a reason for it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.