I am very happy this thread came up I have just bought three 3C24's just wondering if there are any comments on using a VR150 for loading the plate of the driver.
Thanks Nigel
Thanks Nigel
hey-Hey!!!,
If you're going to cook up the plate to getter/de-gas the tubes, do *NOT* leave the cathode hot for long. You want a HV supply ready to heat up the anode( it'll deliver some spectacular arcs in the gassy ones too ), so as not to wreck the cathode with gas. I've saved 4E27's that glowed like 866's at full boogie this way.
cheers,
Douglas
If you're going to cook up the plate to getter/de-gas the tubes, do *NOT* leave the cathode hot for long. You want a HV supply ready to heat up the anode( it'll deliver some spectacular arcs in the gassy ones too ), so as not to wreck the cathode with gas. I've saved 4E27's that glowed like 866's at full boogie this way.
cheers,
Douglas
A quick prototype
I decided to modify a previous prototype board to test a 3C24 in A2 mode. I designed an input amp and dynamic driver using a 6FQ7 with a variable bias adjustment for the 3C24 idle current. I also added a DC filament supply for the 3C24.
With a 3.5K load I can manage about 4 watts output with no feedback. It suffers from dynamic compression after 3 watts but reasonably clean, has an input sensitivity of 800mv RMS for 1-watt output, THD at 1-watt is a bit under 1%. Signal-to-noise ratio is ~80dB referenced to 1-watt output.
The down side is I need to push the cathode current up to 90ma with a plate voltage of +330 volts and a grid voltage around +27 volts. This certainly is pushing the 3C24 beyond it's usual ratings, but much below that and output power suffers.
Regards, KM
I decided to modify a previous prototype board to test a 3C24 in A2 mode. I designed an input amp and dynamic driver using a 6FQ7 with a variable bias adjustment for the 3C24 idle current. I also added a DC filament supply for the 3C24.
With a 3.5K load I can manage about 4 watts output with no feedback. It suffers from dynamic compression after 3 watts but reasonably clean, has an input sensitivity of 800mv RMS for 1-watt output, THD at 1-watt is a bit under 1%. Signal-to-noise ratio is ~80dB referenced to 1-watt output.
The down side is I need to push the cathode current up to 90ma with a plate voltage of +330 volts and a grid voltage around +27 volts. This certainly is pushing the 3C24 beyond it's usual ratings, but much below that and output power suffers.
An externally hosted image should be here but it was not working when we last tested it.
Regards, KM
http://babelfish.yahoo.com/translat...ie/KY-3C24-J.html&lp=ja_en&btnTrUrl=Translate
Some more info about that schematic...
Some more info about that schematic...
Re: Re: A quick prototype
Agreed.... need to get proper heat-sinks for these, just using some heavy gauge solid copper wound around each lead. I'm pushing them hard but you need to get enough positive grid voltage to keep the dynamic driver from dropping off at larger voltage swings. Based on measured voltages in the valve world link, the current has to be much higher (than 75ma) for +30 on the grid. I have 4 of the 3C24 tubes and they all sink more current at the voltages shown.
Still, I think with a bit more work I might get it a bit cleaner at less power dissipation. Could be that some form of feedback might be required to get it clean enough. It doesn't sound bad, but it could be better.
Regards, KM
316a said:
Agreed.... need to get proper heat-sinks for these, just using some heavy gauge solid copper wound around each lead. I'm pushing them hard but you need to get enough positive grid voltage to keep the dynamic driver from dropping off at larger voltage swings. Based on measured voltages in the valve world link, the current has to be much higher (than 75ma) for +30 on the grid. I have 4 of the 3C24 tubes and they all sink more current at the voltages shown.
Still, I think with a bit more work I might get it a bit cleaner at less power dissipation. Could be that some form of feedback might be required to get it clean enough. It doesn't sound bad, but it could be better.
Regards, KM
schematic for 3C24 SE amp
I've done a fair amount of testing with the 3C24 and a dynamic follower circuit (different from Valve World's). I opted for no feedback and have found a pretty narrow operating point for the 3C24 in A2 mode. The voltages listed on the VW schematic shows +30 volts on the grid and 290 volts on the plate. It's pretty ideal for getting 4-watts of power. However, two main differences between findings... first, a 3.5K load was pretty spot on for efficiency. Dropping to 2.5K load resulted is power loss and higher distortion. Second, idle current is quite a bit higher than noted as 75ma. To get +30 on the grid, plate current is closer to 85ma and consistent with four different tubes.
Decreasing the idle current reduces the grid voltage and available signal swing to the dynamic follower. You can't drive the grid to 0-volts as current flow decreases before this and the dynamic follower goes into cut-off. Driving positive has it's limit as well so this circuit topology limits your operating points. Overall performance was decent but not stellar. THD is about 1% at 1-watt and increases close to linear with power, reaching about 3.5% at 4-watts output. Bandwidth was no issue with good iron and within 1dB at 50KHz. There is noticeable dynamic compression starting in around 3-watts and increasing to clipping, which is about 4-watts. Swapping enough tubes around, average gain results in 1-watt output with 775mv input. A simple DC filament supply on the 3C24 resulted in signal-to noise better than 80dB referenced to 1-watt.
Now the burning question... should I build a pair for actual use... it is a pretty tube. Overall performance numbers are akin to a 2A3, albeit with more distortion and over twice the power to lite it up. Is anyone currently listening to an amp with 3C24 outputs?
Regards, KM
I've done a fair amount of testing with the 3C24 and a dynamic follower circuit (different from Valve World's). I opted for no feedback and have found a pretty narrow operating point for the 3C24 in A2 mode. The voltages listed on the VW schematic shows +30 volts on the grid and 290 volts on the plate. It's pretty ideal for getting 4-watts of power. However, two main differences between findings... first, a 3.5K load was pretty spot on for efficiency. Dropping to 2.5K load resulted is power loss and higher distortion. Second, idle current is quite a bit higher than noted as 75ma. To get +30 on the grid, plate current is closer to 85ma and consistent with four different tubes.
Decreasing the idle current reduces the grid voltage and available signal swing to the dynamic follower. You can't drive the grid to 0-volts as current flow decreases before this and the dynamic follower goes into cut-off. Driving positive has it's limit as well so this circuit topology limits your operating points. Overall performance was decent but not stellar. THD is about 1% at 1-watt and increases close to linear with power, reaching about 3.5% at 4-watts output. Bandwidth was no issue with good iron and within 1dB at 50KHz. There is noticeable dynamic compression starting in around 3-watts and increasing to clipping, which is about 4-watts. Swapping enough tubes around, average gain results in 1-watt output with 775mv input. A simple DC filament supply on the 3C24 resulted in signal-to noise better than 80dB referenced to 1-watt.
An externally hosted image should be here but it was not working when we last tested it.
Now the burning question... should I build a pair for actual use... it is a pretty tube. Overall performance numbers are akin to a 2A3, albeit with more distortion and over twice the power to lite it up. Is anyone currently listening to an amp with 3C24 outputs?
Regards, KM
Re: Re: Re: A quick prototype
The easiset way of implementing feedback is to put the output transformer secondary in the cathode circuit . I found with bass reflex speakers the bass was very loose and I had to use cathode feedback on the 3C24 and local feedback from the 3C24 plate to the driver
316A
kmaier said:
The easiset way of implementing feedback is to put the output transformer secondary in the cathode circuit . I found with bass reflex speakers the bass was very loose and I had to use cathode feedback on the 3C24 and local feedback from the 3C24 plate to the driver
316A
Hi 316A,
I implemented a local feedback arrangement a couple nights ago from the 3C24 plate to the cathode of the input amp... it helped clean things up a bit and THD was cut down by a reasonable amount but so was gain by ~6dB. I also thought about using the secondary for cathode feedback but opted instead to fine tune the initial non-feedback design.
The schematic posted has the latest changes and is pretty optimal. Trying to determine an optimal idle current for the 3C24 was of little value, but instead setting the grid voltage to +28.5 volts yields the best linearity and symmetry of the output stage. This results in an idle current ~82ma which is 10% over the published specifications, but this is where the tube performs its best. At 24 watts, plate dissipation is just within the 25-watt rating.
Regards, KM
I implemented a local feedback arrangement a couple nights ago from the 3C24 plate to the cathode of the input amp... it helped clean things up a bit and THD was cut down by a reasonable amount but so was gain by ~6dB. I also thought about using the secondary for cathode feedback but opted instead to fine tune the initial non-feedback design.
The schematic posted has the latest changes and is pretty optimal. Trying to determine an optimal idle current for the 3C24 was of little value, but instead setting the grid voltage to +28.5 volts yields the best linearity and symmetry of the output stage. This results in an idle current ~82ma which is 10% over the published specifications, but this is where the tube performs its best. At 24 watts, plate dissipation is just within the 25-watt rating.
Regards, KM
Re: schematic for 3C24 SE amp
When the 20k pot that sets the bias for the second tube fails, it will likely fail with the wiper open (common failure mode for pots). This will leave the second tube without any DC path to ground on the grid, thus, I would imagine that tube being quite unhappy under these circumstances. I think it would be worthwhile to set up the bias circuit to fail safely. This can be done by connecting the wiper of the pot to one of the ends of the pot (the end that will cause the lowest bias current in the tube should the pot fail). Use the pot as a variable resistor rather than a potentiometer/variostat.
Just a thought...
~ Tom
kmaier said:
When the 20k pot that sets the bias for the second tube fails, it will likely fail with the wiper open (common failure mode for pots). This will leave the second tube without any DC path to ground on the grid, thus, I would imagine that tube being quite unhappy under these circumstances. I think it would be worthwhile to set up the bias circuit to fail safely. This can be done by connecting the wiper of the pot to one of the ends of the pot (the end that will cause the lowest bias current in the tube should the pot fail). Use the pot as a variable resistor rather than a potentiometer/variostat.
Just a thought...
~ Tom
Tom,
Good point.... pots do fail. However, I would be inclined to put a higher value resistor from the wiper pin to ground. This would provide a near zero bias in case of pot failure and idle the output section down low as a result. If it does fail, I'd prefer the output stage go low on current so hopefully someone notices ;-)
I've since changed the schematic a bit to provide some regulation to the bias voltage to minimize drift from AC line voltage fluctuations, so I added this suggestion as well. I've also swapped out the OPT for an Edcor unit... my Dad want's to build these, so working on a version using Edcor Power and OPT and a Hammond choke. Here's the updated schematic.
Regards, KM
Good point.... pots do fail. However, I would be inclined to put a higher value resistor from the wiper pin to ground. This would provide a near zero bias in case of pot failure and idle the output section down low as a result. If it does fail, I'd prefer the output stage go low on current so hopefully someone notices ;-)
I've since changed the schematic a bit to provide some regulation to the bias voltage to minimize drift from AC line voltage fluctuations, so I added this suggestion as well. I've also swapped out the OPT for an Edcor unit... my Dad want's to build these, so working on a version using Edcor Power and OPT and a Hammond choke. Here's the updated schematic.
An externally hosted image should be here but it was not working when we last tested it.
Regards, KM
Can't find information
Could some one please explain what Tossie refers to as dynamic coupling .
Thanks Nigel
Could some one please explain what Tossie refers to as dynamic coupling .
Thanks Nigel
Re: Can't find information
Dynamic-Coupling is simply a term used to define the topology of driving a grid in Class A2 mode without resorting to an expensive inter-stage transformer. I've looked at this topology many times before doing a design based on it. In short, I consider it a cost saving rather than a performance circuit. If you think about operational characteristics, the flaws and limitations become obvious.
How it works is quite simple. Using the driver stage as a cathode follower, the cathode loading, which is usually a simple resistor to ground, is using the grid of the output tube to conduct current. This is where the limitations set in. With the 3C24 grid at ground, a fair amount of current still flows, depending on plate voltage could be 20-30ma. As you are using a cathode follower with the output tube grid as it's load, as you swing negative, the cathode follower (aka dynamic coupler) will go into cutoff before you reach zero on the grid and it's operation becomes non-linear so increased distortion is the result. Also, as you drive the grid positive you can top-out your driver before you ultimately reach maximum current flow in the output stage... or possibly drive the output tube into saturation.
This results in a bit of a balancing act. You need to pick an idle current with sufficient grid voltage that allows enough voltage swing (both positive and negative) that gives you a reasonable range which centers your linear operating conditions for the output stage and driver stage. For the 3C24 I did a lot of testing and found that Tossie's voltages are pretty much optimum. I did my own design and provided some adjustment range. I did not find 2.5K as good load however, 3.5K worked much better... more power, less distortion.
Hope this helps.
Regards, KM
Dynamic-Coupling is simply a term used to define the topology of driving a grid in Class A2 mode without resorting to an expensive inter-stage transformer. I've looked at this topology many times before doing a design based on it. In short, I consider it a cost saving rather than a performance circuit. If you think about operational characteristics, the flaws and limitations become obvious.
How it works is quite simple. Using the driver stage as a cathode follower, the cathode loading, which is usually a simple resistor to ground, is using the grid of the output tube to conduct current. This is where the limitations set in. With the 3C24 grid at ground, a fair amount of current still flows, depending on plate voltage could be 20-30ma. As you are using a cathode follower with the output tube grid as it's load, as you swing negative, the cathode follower (aka dynamic coupler) will go into cutoff before you reach zero on the grid and it's operation becomes non-linear so increased distortion is the result. Also, as you drive the grid positive you can top-out your driver before you ultimately reach maximum current flow in the output stage... or possibly drive the output tube into saturation.
This results in a bit of a balancing act. You need to pick an idle current with sufficient grid voltage that allows enough voltage swing (both positive and negative) that gives you a reasonable range which centers your linear operating conditions for the output stage and driver stage. For the 3C24 I did a lot of testing and found that Tossie's voltages are pretty much optimum. I did my own design and provided some adjustment range. I did not find 2.5K as good load however, 3.5K worked much better... more power, less distortion.
Hope this helps.
Regards, KM
Thanks KM for your reply it gives me some where to start . I received three 3c24's the other day 2 Lewis and one Hentz&Kaufman the only difference I can see is that one has three supports on the anode and the others have two they are a pretty little valve .
Thanks
Nigel
Thanks
Nigel
Hi Nigel,
I'm not certain how many manufacturers were making 3C24 tubes but in addition to the two (brands) you have there was North American Philips (aka Norelco), General Electronics and Eimac... maybe more. I have 3 NOS JAN from Norelco dated April 5th and 6th 1944 in the original boxes and one unmarked rogue which I've had in my possession for 40+ years (it's mate got broken about 35 years ago... sniff). Some brands have a ceramic base and aluminum collar where others have a standard bakelite base. With the amount of heat these can generate under normal conditions (~45 watts), it is critical to ensure adequate airflow and proper heat-sinking of the anode and grid leads. In any case, the Eimac version was thought to be the "creme de la creme" but the H&K ones are also excellent.
From the stock I have, they all work fine and are very close in operational specs to each other... which is very good for tubes made over 65 years ago. I still find having a bias adjustment to be a plus.... you can actually fine tune operation for minimal distortion using an analyzer. The only kicker is that depending on the output power level for the distortion measurement, you have slightly different bias settings for different power levels.
Regards, KM
I'm not certain how many manufacturers were making 3C24 tubes but in addition to the two (brands) you have there was North American Philips (aka Norelco), General Electronics and Eimac... maybe more. I have 3 NOS JAN from Norelco dated April 5th and 6th 1944 in the original boxes and one unmarked rogue which I've had in my possession for 40+ years (it's mate got broken about 35 years ago... sniff). Some brands have a ceramic base and aluminum collar where others have a standard bakelite base. With the amount of heat these can generate under normal conditions (~45 watts), it is critical to ensure adequate airflow and proper heat-sinking of the anode and grid leads. In any case, the Eimac version was thought to be the "creme de la creme" but the H&K ones are also excellent.
From the stock I have, they all work fine and are very close in operational specs to each other... which is very good for tubes made over 65 years ago. I still find having a bias adjustment to be a plus.... you can actually fine tune operation for minimal distortion using an analyzer. The only kicker is that depending on the output power level for the distortion measurement, you have slightly different bias settings for different power levels.
Regards, KM
Mines are H&K / Eimac too ... IMHO General Electronics did not build such tubes (patented) and , as many other refs , just rebranded some batches before implementation in GE machines . Ever tried the HK54 ? ... a fatter 3C24 🙂
Hi Coresta,
No doubt they (General Electronics) could have re-branded them. That was very common between manufacturers. Here's a pic from a guy who has some:
http://www.sphere.bc.ca/test/tubes12/jan-3c24.jpg
Sadly, I have no HK54 tubes.... but do have the spec sheet. These would also work well and as suggested earlier, would be a bump up. I'm also looking at the 35T from Eimac, looks like another upgrade option.
My main reason for adopting some of these older transmitting tubes (besides looking pretty when lit) is their price is still attractive and they should certainly have longer life than a 45 or 2A3. The market prices for 45s and 2A3s has shot up and the actual quality of the remaining tubes has shot down.
Are you also using your 3C24s in an audio amplifier?
Regards, KM
No doubt they (General Electronics) could have re-branded them. That was very common between manufacturers. Here's a pic from a guy who has some:
http://www.sphere.bc.ca/test/tubes12/jan-3c24.jpg
Sadly, I have no HK54 tubes.... but do have the spec sheet. These would also work well and as suggested earlier, would be a bump up. I'm also looking at the 35T from Eimac, looks like another upgrade option.
My main reason for adopting some of these older transmitting tubes (besides looking pretty when lit) is their price is still attractive and they should certainly have longer life than a 45 or 2A3. The market prices for 45s and 2A3s has shot up and the actual quality of the remaining tubes has shot down.
Are you also using your 3C24s in an audio amplifier?
Regards, KM
Not YET KM, but ... in the next few weeks : brass plates chassis were drilled 2 days ago for that purpose 😉 Among the "waiting" candidates for PP heating : 100TH , 830B, 809, 3C24, HK54 , VT127A ! Stay tuned 😀 regards
Pierre
Pierre
What is PP heating? I am pretty sure that VT127A has the same heater requirement as a 250tl. So you may as well add this big tube to your list.
- Status
- Not open for further replies.