I am interested in designing preamp. With no feedback. The normal way is to decuple the cathode resistor with a cap. Thies cap tend to influence the sound and people tent to spend a lot of money for good quality ones. I was thinking if it would be an option to place a battery in the cathode instead, sines the internal is measured in milli ohms. This battery will of course be decupled as well. And I know that the voltages of the battery must be monitored and discharged in a proper manner over time.
I understand you need to find the right tube for a given battery voltage. Has anyone seen or have any experience with sus a setup?
I understand you need to find the right tube for a given battery voltage. Has anyone seen or have any experience with sus a setup?
This post may be of interest Post in thread 'cathode battery bias questions' https://www.diyaudio.com/community/threads/cathode-battery-bias-questions.35284/post-407988
You will probably find more if you search for battery bias, or NIMH.
You will probably find more if you search for battery bias, or NIMH.
Putting the battery in the cathode is not a good idea, you get the cathode's DC current flowing through the battery. The better solution is the battery in series with the grid stopper resistor, positive connected to the grid leak, negative to the grid stopper. In this case there is no DC current trough it (well, extremely low) and the AC signal goes through like it was a wire.
Example: Post in thread 'Battery bias question' https://www.diyaudio.com/community/threads/battery-bias-question.350715/post-6113091
I use a 1:1 input transformer (Lundahl LL1676) feeding the 01A grid in my preamp/DAC. I have a 9V battery between the transformer secondary and ground. Using a lithium battery it has provided excellent sound and very long life (essentially the shelf life of the battery). No DC blocking cap is necessary because the transformer isolates the DC voltage from the signal input. The attached schematic shows 2:1 configuration but I've since rewired it to 1:1 to give more gain in the preamp for weak sources.
Another neat trick is that you can flip the +/- in the secondary to compensate for the single triode inverting the signal phase.
Another neat trick is that you can flip the +/- in the secondary to compensate for the single triode inverting the signal phase.
Attachments
Last edited:
You can also use the above scheme to directly interface a DAC board to use the triode as an output stage. In my case it's a Buffalo IIISE Pro 9028 with DAC outputs directly wired to the transformer primary. Proper selection of the secondary snubber provides a filter to roll off HF hash from the DAC as a bonus.
This is historically first biasing solution. In 20s-30s three batteries were used to operate tubes: A - filament battery; B - high voltage plate battery; C - grid bias battery. The B+ designation for plate supply voltage comes from that time. In my experience, battery grid bias is the best sounding solution compared to other biasing schemes.
I used Li-ion rechargeable batteries for grid bias. There are 2 advantages of these batteries: 1) voltage can be precisely adjusted between 3.5 - 4.1 V by varying degree of charge; 2) voltage of a good battery remains stable for years.
I used Li-ion rechargeable batteries for grid bias. There are 2 advantages of these batteries: 1) voltage can be precisely adjusted between 3.5 - 4.1 V by varying degree of charge; 2) voltage of a good battery remains stable for years.
An modern day alternative is to use one or a couple of LED in series with the cathode.
A limitation of battery bias is the low number of battery voltages that are available.
Some examples I have used:
1.5V
3 V
4.5V
6V
And multiple 9V in series for approximately 45, 60, and 90V.
9V alkaline batteries are over 9.5V.
I have used all of the above, except 1.5V bias.
Input tubes, driver tubes, and output tubes, I have used the following topology (connections of the battery in the circuit)
An example of a good circuit:
A resistor to ground with the top of the resistor, or potentiometer connected to ground, with the wiper connected to the + of the battery, the - of the battery connects to the grid stopper resistor, and the other end of the grid stopper connected directly to the socket grid tab.
The advantages of this topology are:
The only current in the battery is when a very large signal draws grid current, and that mildly 'charges' the battery.
The only steady current in the battery is the extremely low grid current of a good tube.
The battery life is essentially the shelf life of the battery
The tube will never draw large current at power-on of the amplifier, because bias voltage is always present.
What more do you need if you want to use battery bias?
Some examples I have used:
1.5V
3 V
4.5V
6V
And multiple 9V in series for approximately 45, 60, and 90V.
9V alkaline batteries are over 9.5V.
I have used all of the above, except 1.5V bias.
Input tubes, driver tubes, and output tubes, I have used the following topology (connections of the battery in the circuit)
An example of a good circuit:
A resistor to ground with the top of the resistor, or potentiometer connected to ground, with the wiper connected to the + of the battery, the - of the battery connects to the grid stopper resistor, and the other end of the grid stopper connected directly to the socket grid tab.
The advantages of this topology are:
The only current in the battery is when a very large signal draws grid current, and that mildly 'charges' the battery.
The only steady current in the battery is the extremely low grid current of a good tube.
The battery life is essentially the shelf life of the battery
The tube will never draw large current at power-on of the amplifier, because bias voltage is always present.
What more do you need if you want to use battery bias?
Last edited:
Yes 🙂
1.2V is sort of standard? So it can be connected in serial for n x.
.
But another issue is present. I would like to have Your opinion on this:
When Battery connected without the some current source on it, it is connection without trickle charge, no charge at all. And if the C is connected in parallel with the battery leakage will slowly drop battery level?
So this type of battery biasing, fixed, in grid point, for my opinion needs monitoring of battery level.
.
When the battery is connected in cathode, for automatic grid bias, battery has some amout of trickle charging when device is working. And it should be around 1/40 of capacity in mA. The discharge via some leakage element is minimal, and when working battery is on slow trickle charge.
BUT the Cathode current is factor for chosing the proper capacity of battery. 800mA/h for example is OK for tube Io=800ma/40=20mA cathode current or a bit less...
It depends of trickle charge current specified for battery type in general...
.
Anothe thing:
Batery in Cathode is acting as very high capacitor. And in the conjuction with reactive L element in anode, IF it is present, can ruin the LF BW and LF Phase... Also will react with Coupling C @ output.
Acording to spice simulations.
When R load is present behaves very good.
...
Anyway some monitoring of the batery voltage is very welcome. Can be with simple taster and display check with mini instrument inside the box. And with one instrument, we can inspect more bateries with a switch...
417A/5842 tube works very well with cathode bias of 1.2~1.4Vdc.
Johnny
Johnny
You are overcomplicating things.
Batteries usually perform worse than an LED. Simple as that.
Despite what you might believe, batteries are noisy. A standard alkaline battery has a noise floor of about 2uv. A Nicad is so low that most test gear will struggle to measure it. But this only works when there is very little current flow. As current begins to rise, the noise increases. Keep in mind that a basic resistor and capacitor combo has virtually no noise contribution (the noise is well below the noise floor) so long as you have good airflow around the circuitry.
Compare that to red LED that could be as low as 0.5uv and you see why people are telling you to use LED autobias over battery bias.
So then why do people talk about battery bias? Simple. Batteries are an excellent option for grid biasing. Some people feel like grid biasing is superior to autobias. Sometimes it's super convenient to use batteries to grid bias an output tube in order to keep the B+ lower. Whatever the reason, batteries are a very handy solution in that instance.
You can obviously try battery autobias for yourself, but a lot of people have already tried and found LED autobias MUCH better. Myself included.
People usually do not try at all but just read that someone claim to be better, by subjective reason?
I didn't say that battery is better than led, BUT question in the topic was about batteries
not about LEDs
And then suddenly - led is better...
.
I was talking about NiMh batteries
To Tjj226: I have seen circuits with a LED in the cathode before, but I have trouble finding a LED with a low dynamic resistance. Do you have any suggestions for what type of LED would have a low dynamic resistance that would fit a 6ns7 tube with an operating point at approx. 15mA
The topic changed to LEDs because you are specifically talking about autobias. Everyone is trying to guide you away from using batteries in an autobias configuration.
But like I said, you are more than welcome to try it out. You might find that you like it or that it works for you in your specific circuit.
LEDs will always have a couple ohms of dynamic impedance. But you can always get tricky and parallel some LEDs. You will reduce your dynamic impedance and improve the noise.
A few green LEDs in parallel actually works surprisingly well because even though they are noisier than a red LED, you don't need to put as many in series and are still low noise enough that paralleling a few of them gets you below the noise level of a battery. Hint hint wink wink.
You can also get tricky and use filament bias on an indirectly heated tube. It is a pain in the butt because you have to make a DC supply for your filaments, but if you do it right, you can further improve your noise floor an use a single resistor for biasing.
Everyone in this topic really needs to spend some time staring at this web page.
https://www.mvaudiolabs.com/diy/modern-jfet-noise-measurements/#led
A bit of topic but still related.
Do any of you have any experience with starving the heat filament in order to adjust to the desired Bias for a tube. My thoughts are as following. If this is done correctly it would be possible to let the tube work at 0 VGS and derby there will be no need for a cathode resistor or external bias. But adjust the bias with the voltages for the heat filament. I do realize that you can’t reduce the voltage for the filament to any point, but I am thinking of a reduction of 10-20%.
Any thoughts on this.
There is another benefit as lowering the filament voltages seems to reduce the distortion. Here is a small article about that:http://diyaudioprojects.com/mirror/members.aol.com/sbench102/dht.html
Do any of you have any experience with starving the heat filament in order to adjust to the desired Bias for a tube. My thoughts are as following. If this is done correctly it would be possible to let the tube work at 0 VGS and derby there will be no need for a cathode resistor or external bias. But adjust the bias with the voltages for the heat filament. I do realize that you can’t reduce the voltage for the filament to any point, but I am thinking of a reduction of 10-20%.
Any thoughts on this.
There is another benefit as lowering the filament voltages seems to reduce the distortion. Here is a small article about that:http://diyaudioprojects.com/mirror/members.aol.com/sbench102/dht.html
Starving the filament may be beneficial. You have to experiment with it to see if it benefits your design.
But you can't bias the tube by fiddling with the filament voltage/current. Tube bias specifically refers to the voltage difference between the grid and the cathode. The filament voltage could be 2,3,5,12,74 volts and the grid won't care. The grid only cares about the voltage the cathode is referenced to. So if the cathode is tied to ground, the grid only sees the cathode potential at 0v regardless of the filament voltage.