Hello,
I got the answer from Dave Slagle in search for a kind of " universal choke "" to be used for solid state designs with power supplies that should supply some serious current.
It would be nice if a single double coil choke could be used like the one in the attachment used in a 19eighties supply for a solid state preamp in France.
Not all the info give by Dave would be necessary to know for the people interested in this thread.
BUT if you still are in the proces of buying your parts it could be the right moment to try a choke input because that one will need a power transformer with higher secundairy voltage.
VA rating can be similar because true choke input will give the transformer an easier task
Greetings, Eduard
The rough formula for critical inductance required for a choke input supply is Lcritical=V/I with the I being in milliamps. So a typical tube supply ding 200V @ 50ma requires 4hy of inductance and it is generally good practice to double that number to assure proper choke input operation. So for a 35V @ 5000ma supply Lcrit becomes 7mh and dropping that supply to 1500ma requires 23mhy which makes 100mhy really overkill.
I do not see any benefits from a swinging choke because they are designed to provide choke input filtering at low currents and then as current demand increases the choke saturates slowly transitioning the filter to cap input which increases output voltage to help counteract the voltage losses from increased currents.
> So minimum 5A, DCR 1ohm, 100mH minimum, 2 coil designs to make it possible to change current rating/ DCR and mH specs, usable for LC and CLC.,swinging or not swinging?
> Am i asking to much?
the only thing that is not possilble to change is the swinging vs. not-swinging since that is dertermined by the gap size which will be set in varnish when the choke is finished.
chances are your best approach is to get a custom design done by a local winder since they will be rather large. I can spec the design for you and There is a place here in the states that can do it but they will weigh about 3-4kg each so carriage will be a consideration.
I got the answer from Dave Slagle in search for a kind of " universal choke "" to be used for solid state designs with power supplies that should supply some serious current.
It would be nice if a single double coil choke could be used like the one in the attachment used in a 19eighties supply for a solid state preamp in France.
Not all the info give by Dave would be necessary to know for the people interested in this thread.
BUT if you still are in the proces of buying your parts it could be the right moment to try a choke input because that one will need a power transformer with higher secundairy voltage.
VA rating can be similar because true choke input will give the transformer an easier task
Greetings, Eduard
The rough formula for critical inductance required for a choke input supply is Lcritical=V/I with the I being in milliamps. So a typical tube supply ding 200V @ 50ma requires 4hy of inductance and it is generally good practice to double that number to assure proper choke input operation. So for a 35V @ 5000ma supply Lcrit becomes 7mh and dropping that supply to 1500ma requires 23mhy which makes 100mhy really overkill.
I do not see any benefits from a swinging choke because they are designed to provide choke input filtering at low currents and then as current demand increases the choke saturates slowly transitioning the filter to cap input which increases output voltage to help counteract the voltage losses from increased currents.
> So minimum 5A, DCR 1ohm, 100mH minimum, 2 coil designs to make it possible to change current rating/ DCR and mH specs, usable for LC and CLC.,swinging or not swinging?
> Am i asking to much?
the only thing that is not possilble to change is the swinging vs. not-swinging since that is dertermined by the gap size which will be set in varnish when the choke is finished.
chances are your best approach is to get a custom design done by a local winder since they will be rather large. I can spec the design for you and There is a place here in the states that can do it but they will weigh about 3-4kg each so carriage will be a consideration.
Attachments
Hello,
Because the current is so low it will easy to make a true choke input. The choke in the kaneda preamp is not constructed in the same way like the lundahl chokes. I remember asking lundahl if i can use their chokes the same like the chokes in the kaneda. He told me only if the current in both coils is the same.
So maybe you will have to use two different chokes with the two coils in series or in parallel
Like you can read in my last message it depends on how much current is running through the circuit and how much Henries you want. More Henries will mean less bleeder current needed
The kaneda takes about 50mA. 35÷50= 0,7Henry if you use a bleeder resistor that will take 10 mA you need only 3.5 henries without circuit connected. THIS is important if you dont want the voltage on the caps go up to high.
You can ask Tho here on this thread also from Vietnam to give you some advice. He is the choke input expert in Vietnam.
I do have an original pair of chokes used in the kaneda but i think making a choke input with the right value will work better.
Greetings, Eduard
Because the current is so low it will easy to make a true choke input. The choke in the kaneda preamp is not constructed in the same way like the lundahl chokes. I remember asking lundahl if i can use their chokes the same like the chokes in the kaneda. He told me only if the current in both coils is the same.
So maybe you will have to use two different chokes with the two coils in series or in parallel
Like you can read in my last message it depends on how much current is running through the circuit and how much Henries you want. More Henries will mean less bleeder current needed
The kaneda takes about 50mA. 35÷50= 0,7Henry if you use a bleeder resistor that will take 10 mA you need only 3.5 henries without circuit connected. THIS is important if you dont want the voltage on the caps go up to high.
You can ask Tho here on this thread also from Vietnam to give you some advice. He is the choke input expert in Vietnam.
I do have an original pair of chokes used in the kaneda but i think making a choke input with the right value will work better.
Greetings, Eduard
Look closer that sch you will see the advances our clever man created.
If we use 2 chokes, they must be very large to equal to dual choke in Rdc & L. There are same advances of that way but my E. too bad to talk about it.
You could use any choke in that schm if they have same Rdc and could stand with that current.
Tho is my friend and we some time discuss about choke.
We, diy community in VN, all love him.
If we use 2 chokes, they must be very large to equal to dual choke in Rdc & L. There are same advances of that way but my E. too bad to talk about it.
You could use any choke in that schm if they have same Rdc and could stand with that current.
Tho is my friend and we some time discuss about choke.
We, diy community in VN, all love him.
Pass DIY Addict
Joined 2000
Paid Member
Appreciate the heads up, but this is what makes 90% of the comments in this thread irrelevant to a Class-A, solid state, Pass design. For example, an Aleph-J runs on 24v rails and each channel burns 100w. That's 4A draw per channel, 2A per pos rail and 2A per neg rail. If you make a dual mono power supply, you're looking at a need for 4 chokes, each needing spec near 4A (2x quiescent current draw) for long-term safety factor. If you make a shared power supply for two channels, then you have 4A pos rail current draw and 4A neg rail current draw. This means chokes appropriate to this design need to handle closer to 6-8A full-time draw (8A-10A would be more ideal). The M2 and F4 amps draw somewhat less power (~80wpc), but 3A chokes are still not ideal for this task. I'm not really a fan of engineering things to operate on the "fringe" of the Safe Operating Area for high-current electronics.
Even these "beefy" chokes are too puny...
hi guys,
This is an interesting topic. As most DIYers, I always try to find ways to experiment and improve my audio system in every way I can. CLC is something that I always want to try out to see if it really makes a big difference in sound quality compared to CRC model.
Following this topic closely and consulting with Eric offline, I made a leap of faith to purchase to buy this choke ERSE Super Q 3.0mH 16 AWG 500W Inductor Crossover Coil
ERSE Super Q 3.0mH 16 AWG 500W Inductor Crossover Coil
3.0mH 16 AWG 500W Inductor, Rdc = 0.178 ohms
After playing around with the CRC and CLC with my DIY amp, I have some interesting results to share that might help others or also give someone more motivation to try it out
For the experiment, I use my DIY Aleph J with Antek 500VA 18V-0-18V and point-point soldering CRC filter stage (32 mf - 0.156 ohm - 30mf) I just used whatever caps left on the shelf for this PS filter stage. I use DMM Fluke 179 and measure Vac for the AC ripple
My CRC is 32 mf - 0.156 ohm - 30mf
AC ripple measured at the first C (after rectifier) is 260 mVAC
AC ripple measured at the last C is around 94.5 mV AC
V+ supply for each channel = 21.30 Vdc
V- supply for each channel = 21.28 Vdc
Current draw across R is 0.58V / 0.156ohm = 3.7A for V+ supply
Replaced R by L: CLC - 32 mf -(3mH, 0.178 ohms)- 30mf
AC ripple measured at the first C (after rectifier) is 260 mVAC
AC ripple measured at the last C is around 9.7 mV AC
V+ supply for each channel = 21.34 Vdc
V- supply for each channel = 21.30 Vdc
Current draw across L is 0.7V / 0.178ohm = 3.932A for V+ supply
Observation
- I don't have access to the oscilloscope at home to see the AC ripple waveform change from C to L. However, the AC measurement on Fluke shows a HUGE drop x9.74 in AC ripple going form CRC (94.5mV) to CLC (9.7mV)
- CLC is releasing more heat than CRC. The heatsink of my Alelph J is getting hot fast. I have to use the fan to cool it down. This could explain by more current is drawn across L than R
The most important question of all: How does it sound ? Any differences?
- I need to spend more time going through different music genre to tell but my initial impression is not so much! Even though, the measurement shows a huge improvement in AC ripple, but i did not notice dramatic differences in vocal/instrumental/soundstage just yet. Again, I need to spend more time to make any conclusion
- I feel like going from single power supply to dual mono PS makes more impact in SQ than going from CRC to CLC. When I switched from single to dual mono PS in my F4, I immediately hear the differences in channel separation, deeper soundstage, bass and high note. I do not hear it with CRC -> CLC just yet. Maybe, if I spend enough time listening, i might be able to hear the sound more focus and clearer now that AC ripple goes much lower (we're talking about x10 here)
Further improvement
- Maybe, I could try to improve grounding using big cooper ground plane and using bigger cap to further reduce AC ripple
- Maybe CLCLC model next!
Here's some photos of my measurement going from C to L. Notice the change in DMM
CRC
CLC
Here's another photos of my Dual mono F4 using CRC.
This is an interesting topic. As most DIYers, I always try to find ways to experiment and improve my audio system in every way I can. CLC is something that I always want to try out to see if it really makes a big difference in sound quality compared to CRC model.
Following this topic closely and consulting with Eric offline, I made a leap of faith to purchase to buy this choke ERSE Super Q 3.0mH 16 AWG 500W Inductor Crossover Coil
ERSE Super Q 3.0mH 16 AWG 500W Inductor Crossover Coil
3.0mH 16 AWG 500W Inductor, Rdc = 0.178 ohms
After playing around with the CRC and CLC with my DIY amp, I have some interesting results to share that might help others or also give someone more motivation to try it out
For the experiment, I use my DIY Aleph J with Antek 500VA 18V-0-18V and point-point soldering CRC filter stage (32 mf - 0.156 ohm - 30mf) I just used whatever caps left on the shelf for this PS filter stage. I use DMM Fluke 179 and measure Vac for the AC ripple
My CRC is 32 mf - 0.156 ohm - 30mf
AC ripple measured at the first C (after rectifier) is 260 mVAC
AC ripple measured at the last C is around 94.5 mV AC
V+ supply for each channel = 21.30 Vdc
V- supply for each channel = 21.28 Vdc
Current draw across R is 0.58V / 0.156ohm = 3.7A for V+ supply
Replaced R by L: CLC - 32 mf -(3mH, 0.178 ohms)- 30mf
AC ripple measured at the first C (after rectifier) is 260 mVAC
AC ripple measured at the last C is around 9.7 mV AC
V+ supply for each channel = 21.34 Vdc
V- supply for each channel = 21.30 Vdc
Current draw across L is 0.7V / 0.178ohm = 3.932A for V+ supply
Observation
- I don't have access to the oscilloscope at home to see the AC ripple waveform change from C to L. However, the AC measurement on Fluke shows a HUGE drop x9.74 in AC ripple going form CRC (94.5mV) to CLC (9.7mV)
- CLC is releasing more heat than CRC. The heatsink of my Alelph J is getting hot fast. I have to use the fan to cool it down. This could explain by more current is drawn across L than R
The most important question of all: How does it sound ? Any differences?
- I need to spend more time going through different music genre to tell but my initial impression is not so much! Even though, the measurement shows a huge improvement in AC ripple, but i did not notice dramatic differences in vocal/instrumental/soundstage just yet. Again, I need to spend more time to make any conclusion
- I feel like going from single power supply to dual mono PS makes more impact in SQ than going from CRC to CLC. When I switched from single to dual mono PS in my F4, I immediately hear the differences in channel separation, deeper soundstage, bass and high note. I do not hear it with CRC -> CLC just yet. Maybe, if I spend enough time listening, i might be able to hear the sound more focus and clearer now that AC ripple goes much lower (we're talking about x10 here)
Further improvement
- Maybe, I could try to improve grounding using big cooper ground plane and using bigger cap to further reduce AC ripple
- Maybe CLCLC model next!
Here's some photos of my measurement going from C to L. Notice the change in DMM
CRC
CLC
Here's another photos of my Dual mono F4 using CRC.
Last edited:
I did in my post. The coil is drawing a bit more current and Rdc is higher than R in CRC network. P = I^2 * R. Higher power -> more heat.
My only 2 cent
Tom
My only 2 cent
Tom
V rail is a tad higher but not significant going from CLC to CRC
I is higher from CLC to CRC
Current draw across L is 0.7V / 0.178ohm = 3.932A for V+ supply
Current draw across R is 0.58V / 0.156ohm = 3.7A for V+ supply
You should try it for yourself and report back. Maybe your experience is different than mine. Heat is something hard to quantify since i don't have the thermometer to precisely tell the difference. I just use my hand for the test most of the time. One day, I should buy a proper instrument for thermal test. Aleph J is a hot amp regardless. I need more heatsink for my current DIY chassis
I is higher from CLC to CRC
Current draw across L is 0.7V / 0.178ohm = 3.932A for V+ supply
Current draw across R is 0.58V / 0.156ohm = 3.7A for V+ supply
You should try it for yourself and report back. Maybe your experience is different than mine. Heat is something hard to quantify since i don't have the thermometer to precisely tell the difference. I just use my hand for the test most of the time. One day, I should buy a proper instrument for thermal test. Aleph J is a hot amp regardless. I need more heatsink for my current DIY chassis
Last edited:
William, What is your take on why there is an increase in current supply when transistor biasing is not changed ?
Here's my last take on this with my limited knowledge. Please take it with a pinch of salt. Maybe other experts can have better explanation.
At the output, you have 2 components that is drawing current, source degeneration resistor and CCS that is controlling how much current sinking through the output transistors. These are not changed apparently.
From the power supply filter ckt, there is also R that is drawing current. If R changes, current could change as well. In this case, I'm changing from CRC to CLC with Rdc of coil that is higher from total resistance in CRC ckt. Thus slight change in total current drawn.
Tom
Here's my last take on this with my limited knowledge. Please take it with a pinch of salt. Maybe other experts can have better explanation.
At the output, you have 2 components that is drawing current, source degeneration resistor and CCS that is controlling how much current sinking through the output transistors. These are not changed apparently.
From the power supply filter ckt, there is also R that is drawing current. If R changes, current could change as well. In this case, I'm changing from CRC to CLC with Rdc of coil that is higher from total resistance in CRC ckt. Thus slight change in total current drawn.
Tom
Last edited:
Pass DIY Addict
Joined 2000
Paid Member
Interesting results for sure! A 9X reduction in ripple at the second power supply cap is impressive with L vs R in the circuit. I suspect the coil will also attenuate higher frequency noise better than a resistor as well, resulting in overall lower noise floor and less "contamination" from the AC mains.
Hmmm... I wonder if the difference in reported bias level is simply a difference in resistance between what the manufacturer of the coil publicizes for specs and the *actual* 120Hz resistance of the coil. Perhaps the resistance of the coil is further away from 0R178 than we think.
I guess a more accurate way to measure bias is to measure the voltage drop across the source resistors for the output mosfets and add them up. I doubt that this has changed.
Hmmm... I wonder if the difference in reported bias level is simply a difference in resistance between what the manufacturer of the coil publicizes for specs and the *actual* 120Hz resistance of the coil. Perhaps the resistance of the coil is further away from 0R178 than we think.
I guess a more accurate way to measure bias is to measure the voltage drop across the source resistors for the output mosfets and add them up. I doubt that this has changed.
Last edited:
Hi Tom,
sorry for the confusion. I seem to be having problems stating what I want to say.
I’ll try again:
Your amps bias hasn’t changed
The voltages haven’t changed
The current running through supply and amp hasn’t changed
So total dissipation hasn’t changed
So how can it get any hotter?
I’ve tried LC, CLC, CRC in my Aleph5 with big influence on supply ripple but not much if any on sound quality. At the moment it’s CRC.
In my Aleph-X I’m using CLC with torobars.
William
sorry for the confusion. I seem to be having problems stating what I want to say.
I’ll try again:
Your amps bias hasn’t changed
The voltages haven’t changed
The current running through supply and amp hasn’t changed
So total dissipation hasn’t changed
So how can it get any hotter?
I’ve tried LC, CLC, CRC in my Aleph5 with big influence on supply ripple but not much if any on sound quality. At the moment it’s CRC.
In my Aleph-X I’m using CLC with torobars.
William
Eric,
That's right 9x-10x is a huge drop. Maybe if I have time, i will go back and measure current drop across source resistor when using CRC and CLC to know for sure if biasing is changing if any. I doubt not but only for sure if measured
William,
I got your point. As I mentioned earlier, i don't have an accurate instrument to quantify the heat/temperature. The measurement shows a bit increase in current draw across the coil vs R. I play music for a bit and feel that the amp is getting hot faster than it's used to but not crazy hot. That's all for my observation.
Btw, I was hoping for some real sonic improvement but not much so far. However, I will continue to use CLC model going forward now that I know the AC ripple is way lower than CRC.
That's right 9x-10x is a huge drop. Maybe if I have time, i will go back and measure current drop across source resistor when using CRC and CLC to know for sure if biasing is changing if any. I doubt not but only for sure if measured
William,
I got your point. As I mentioned earlier, i don't have an accurate instrument to quantify the heat/temperature. The measurement shows a bit increase in current draw across the coil vs R. I play music for a bit and feel that the amp is getting hot faster than it's used to but not crazy hot. That's all for my observation.
Btw, I was hoping for some real sonic improvement but not much so far. However, I will continue to use CLC model going forward now that I know the AC ripple is way lower than CRC.
Hello,
CRC and CLC will sound completely different if you use the right kind of choke.
I think 3 mH is much to low.
The French in the past used 200 mH 0,7 ohm dcr in some of their designs.
Take a look at the Hammond 195T5 100mH 5A 0,64 ohm. YES they are big, heavy and not cheap. As long if you will not use them for choke input ( they will probably get noisy) it will be a good investment.
Available at Mouser and free shipping.
Greetings, Eduard
CRC and CLC will sound completely different if you use the right kind of choke.
I think 3 mH is much to low.
The French in the past used 200 mH 0,7 ohm dcr in some of their designs.
Take a look at the Hammond 195T5 100mH 5A 0,64 ohm. YES they are big, heavy and not cheap. As long if you will not use them for choke input ( they will probably get noisy) it will be a good investment.
Available at Mouser and free shipping.
Greetings, Eduard
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.