They’re 100 ohm - they simply don’t draw that much wattage. So what if they are only good for 2 watts that way. Mounting metal cased ones down to wood is probably better than leaving a white sand box variety dangling. Especially if you intend on leaving it that way.
I’ve done it too - but not without obvious distortion. I was assuming it was to be kept “clean” when comparing speakers.
When comparing amps I’ll drive them way into clipping. Most times it’s the only time you hear obvious differences. You’ll want a speaker that can take the abuse too.
When comparing amps I’ll drive them way into clipping. Most times it’s the only time you hear obvious differences. You’ll want a speaker that can take the abuse too.
I wish to do something similar, similar but not the same.
I wish to wire two transformer primaries to the output of an SE tube amp, and switch one lead of the output through a relay to the speaker + lead, with both common leads wired to the speaker.
The relay will be Break before Make.
I am looking to find the highest value of resistance which wired directly across the OPT secondary (before the relay contacts) will protect the OPT.
I have seen 100R recommended. If the speaker impedance swings up to 32 ohms at high frequencies, the 100R would seem to become significant.
Does anyone have experience with greater values" 1K?
I wish to wire two transformer primaries to the output of an SE tube amp, and switch one lead of the output through a relay to the speaker + lead, with both common leads wired to the speaker.
The relay will be Break before Make.
I am looking to find the highest value of resistance which wired directly across the OPT secondary (before the relay contacts) will protect the OPT.
I have seen 100R recommended. If the speaker impedance swings up to 32 ohms at high frequencies, the 100R would seem to become significant.
Does anyone have experience with greater values" 1K?
Spark gaps have been used extensively in the past to protect audio output transformers and transmitter modulation transformers.
I remember seeing a lot of tube sockets, tube bases and transformers ruined by arc burns, indeed sometimes a tube base could be saved with a saw cut through the burn. Many tubes with a saw cut go on and last for years.
I have seen many amplifiers with spark gap protection showing obvious signs of spark gap erosion with the rest of the circuit functioning normally.
The spark gap was placed across the transformer winding never from plate to ground as once lit it would continue to burn on the DC supply. When the spark gap fires its resistance will drop to low ohms or less so a load resistor must be placed in series with the spark gap. I remember a guitar amp that had a spark gap made on a tag strip; it seemed to do the job.
From memory I think the gap is about 1mm for 3kV.
Spark gaps have no parasitic load and very low capacitance.
https://www.researchgate.net/figure...-as-a-Function-of-Pressure-for_fig6_267576105
https://www.diyaudio.com/community/threads/spark-gaps-on-output-transformers.183446/
I have enclosed a schematic from Babani #12 book on amplifiers.
I found this description of a Philips amplifier : Philips 979. Five input (4x MIC, 1x PU) channel PA amplifier. 2 MIC inputs to 2x EF86. 2 MIC inputs to 12AX7. 6N8 mixer. Half 12AX7 tone make-up, with feedback to part-cathode. Half 12AX7 cathodyne PI. 6CA7 fixed bias PP, with spark-gap protection. 2x 5U4G Doubler HT. 6V4 for negative bias supply. Circa 1955-6.
I remember seeing a lot of tube sockets, tube bases and transformers ruined by arc burns, indeed sometimes a tube base could be saved with a saw cut through the burn. Many tubes with a saw cut go on and last for years.
I have seen many amplifiers with spark gap protection showing obvious signs of spark gap erosion with the rest of the circuit functioning normally.
The spark gap was placed across the transformer winding never from plate to ground as once lit it would continue to burn on the DC supply. When the spark gap fires its resistance will drop to low ohms or less so a load resistor must be placed in series with the spark gap. I remember a guitar amp that had a spark gap made on a tag strip; it seemed to do the job.
From memory I think the gap is about 1mm for 3kV.
Spark gaps have no parasitic load and very low capacitance.
https://www.researchgate.net/figure...-as-a-Function-of-Pressure-for_fig6_267576105
https://www.diyaudio.com/community/threads/spark-gaps-on-output-transformers.183446/
I have enclosed a schematic from Babani #12 book on amplifiers.
I found this description of a Philips amplifier : Philips 979. Five input (4x MIC, 1x PU) channel PA amplifier. 2 MIC inputs to 2x EF86. 2 MIC inputs to 12AX7. 6N8 mixer. Half 12AX7 tone make-up, with feedback to part-cathode. Half 12AX7 cathodyne PI. 6CA7 fixed bias PP, with spark-gap protection. 2x 5U4G Doubler HT. 6V4 for negative bias supply. Circa 1955-6.
Last edited:
My spark gaps are 230VAC type, so too low voltage for 400V B+.
I do have some 1.5KE 440A transorbs, so I will use them back to back across the primary, with 1K across the secondary.
I have a push-push button switch to use for switching the relay. I can't tell which way it is by feel (on or off) so the test should be truly double blind if a second person operates the switch for the listener.
Thanks
I do have some 1.5KE 440A transorbs, so I will use them back to back across the primary, with 1K across the secondary.
I have a push-push button switch to use for switching the relay. I can't tell which way it is by feel (on or off) so the test should be truly double blind if a second person operates the switch for the listener.
Thanks
Hi TheGimp
That sounds like a good plan.
Transorbs can have quite high capacitance, often greater than 500pF, this may have an effect on the performance of the amplifier. It will be interesing to hear the result of your transorb test.
The operation of a transorb and a spark gap are very different, a transorb will clamp the voltage while a spark gap will function more like a switch hence the need for a series resistor.
I remember viewing a spark gap as a guitar was strummed with no speaker load on the amplifier, it sparked away each time the instrument was strummed; with the load connected it could not be made to spark.
The next time I use a spark gap I will try using tungsten TIG welding electrodes as the electrode material.
Thanks for that, the author loves metal oxide varistors MOV's. You need to be careful with the design I've seen MOV's fail.
here is another interesting link.
https://pearl-hifi.com/06_Lit_Archive/15_Mfrs_Publications/GEC_UK/GEC-UK_AF_Amplifier_Design.pdf
Was that in the output transformer protection application, or perhaps in the very common mains ac primary circuit application?
They were in the very common AC primary circuit application and most were not suitably rated; it does however serve to remind us to design things correctly.
BTW I have nothing against MOV's, transorbs or zener diodes as protection devices, when correctly implemented all are good.
A microprocessor implemented warning and relay activated dummy load is another option.
The micro could detect correct load restoration and disconnect the dummy load restoring normal operation.
Transorbs would still be required to cover maybe 500 milliseconds.
If microprocessor monitoring was in place it could display the actual load impedance and system health. I have been wondering if using an e-paper label would be a good idea it could display a logo when the amplifier was off and magically change when a fault occurred.
Maybe an Arduino board could be used so it would be easy and cheap for all to implement.
Ken K
I am switching without turning the input to zero. Big stress possibly.
So far several people preferred one setting to the other stating it sounded better defined, clearer, more detailed.
There was a 1dB difference in the two outputs. I corrected that and will repeat the test.
So far several people preferred one setting to the other stating it sounded better defined, clearer, more detailed.
There was a 1dB difference in the two outputs. I corrected that and will repeat the test.
Good point Tim, if I was just looking to detect open circuit output I would look at the low z output (speaker) measuring the voltage and current to the load this will enable the micro to calculate the load impedance and absence of load hence switching in a dummy load, one could also calculate the power output and amplifier clipping. The voltage divider from the speaker output would have no bandwidth issues. As you inferred a buffer may be needed for the anode measurement if a high bandwidth is needed.
If a micro was in place it could monitor, B+ voltage and current, bias voltage, heater voltage and current, the is bandwidth not so important for these so the ADC input capacitance should be fine. Some of the hall effect current sensors would be good on the low z output to the speaker as no ground loops, complete galvanic isolation, and can measure AC, 5V supply and output suit micro controllers.
https://docs.rs-online.com/9661/0900766b813b538d.pdf
I see the allegro ACS723LLCTR-05AB-T has a span of +/- 5A and has a dielectric strength of 2.4 KVrms and a working voltage up to 420 VDC, this device could measure B+ current in larger amps it output is 400mV/A. +/- 5A would be ok for the amplifier current in many cases.
If one used a mico it could also control a B+ delay on start up, and to reduce the tube current in the absence of audio for a given time period or even auto off.
Tim I mentioned Arduino as they are cheap and plentiful, I have purchased quite a few and never used one as I always seem to design my own PCBs and use PIC microcontrollers.
Maybe if any person wants to undertake an Arduino or ESP32 project we could have a special thread in this group, there are a lot of smart people in this group maybe a group effort, it may be possible to use all cheap eBay boards with few external components.
Ken
Are you suggesting to rectify and RC detect an ac signal to present a dc voltage to a Pic or Arduino analog input, for both speaker side voltage and current signals and then apply code logic to continuously apply a test that the signals represent a suitably loaded OPT secondary winding, no matter what impedance is being presented by any speaker, or the frequency content being passed, or the range of signal amplitudes anticipated?
Hi Tim
You are correct in assuming the voltage and current waveforms could be rectified and then presented as a DC signal to a microprocessor board.
My thought on this idea was to use an Arduino as the sketch (Arduino program) could be written or modified by many thousands of people capable of doing so. If a dedicated microcontroller using a PIC was made I'm guessing less people would be able to easily edit the program.
I was trying to avoid designing a special PCB.
It may be possible using a micro with a fast ADC to analyse the waveform negating the need for much additional circuitry. I have not fully investigated but a ESP32 might do the job and Arduino software and sketches can be used to program them.
Indeed it may be possible to include a real-time spectrum analyser something like this.
https://www.waitingforfriday.com/?p=325
Some ESP32 board have blue tooth capability making other features possible.
Below is a quick spice sim of how voltage and current data could be simply presented to such a device.
If using a small Arduino or PIC PCB one could make up a small external PCB that had a power supply for the micro, current sensor, and relays, the relays could also be included on this board. The board could also include the precision audio full rectifiers and filters for current and voltage waveforms, maybe even one of those true RMS power chips.
I will now ramble on with a story. Years ago I did some work a a race track, we used a rack of Grampian 100 watt tube amplifiers with 100 volt lines feeding large exponential horn speakers c/w Tannoy drivers, each amp had a moving coil AC voltmeter on it's output and a moving coil AC ammeter on its output. We could tell when a amplifier lost a speaker or when we had a short on a line (high current low volts and maybe red plating KT66's). People would set up tents and sometimes drive a peg through a speaker wire, pull the peg, meters back to normal, anode stops glowing. We tried solid state amps but they kept blowing up from the abuse.
We are looking at replacing the meters and eyeballs with some automatic modern technology.
Regarding the Grampian amplifiers we purchased GE KT66 tubes by the carton and changed them all once a year, all the old ones were thrown in the trash. 🙁 < 250 hours old!
You are correct in assuming the voltage and current waveforms could be rectified and then presented as a DC signal to a microprocessor board.
My thought on this idea was to use an Arduino as the sketch (Arduino program) could be written or modified by many thousands of people capable of doing so. If a dedicated microcontroller using a PIC was made I'm guessing less people would be able to easily edit the program.
I was trying to avoid designing a special PCB.
It may be possible using a micro with a fast ADC to analyse the waveform negating the need for much additional circuitry. I have not fully investigated but a ESP32 might do the job and Arduino software and sketches can be used to program them.
Indeed it may be possible to include a real-time spectrum analyser something like this.
https://www.waitingforfriday.com/?p=325
Some ESP32 board have blue tooth capability making other features possible.
Below is a quick spice sim of how voltage and current data could be simply presented to such a device.
If using a small Arduino or PIC PCB one could make up a small external PCB that had a power supply for the micro, current sensor, and relays, the relays could also be included on this board. The board could also include the precision audio full rectifiers and filters for current and voltage waveforms, maybe even one of those true RMS power chips.
I will now ramble on with a story. Years ago I did some work a a race track, we used a rack of Grampian 100 watt tube amplifiers with 100 volt lines feeding large exponential horn speakers c/w Tannoy drivers, each amp had a moving coil AC voltmeter on it's output and a moving coil AC ammeter on its output. We could tell when a amplifier lost a speaker or when we had a short on a line (high current low volts and maybe red plating KT66's). People would set up tents and sometimes drive a peg through a speaker wire, pull the peg, meters back to normal, anode stops glowing. We tried solid state amps but they kept blowing up from the abuse.
We are looking at replacing the meters and eyeballs with some automatic modern technology.
Regarding the Grampian amplifiers we purchased GE KT66 tubes by the carton and changed them all once a year, all the old ones were thrown in the trash. 🙁 < 250 hours old!
Last edited:
Trying to limit violent voltage spikes AFTER the fact will always be to fight fire after it has broken out.
To prevent fire will prevent damage instead.
Preventing overvoltage in tubeamps is known knowledge. It's cheap and it won't involve any
active components. A simple resistor or zobel network is all that is needed and it will prevent
damage.
An arduino however can be of great assistance in other areas. Two of them is to monitor cathod current
( voltage across a resistor) and shut down the amp while keeping the reason why and whick tube that is
the problem. Another function is to assist in powering up and down , preventing rapis off/on events,
and ensure proper timing at power on and off.
But acting on sudden flash is not where an arduino can be useful. (inho)
To prevent fire will prevent damage instead.
Preventing overvoltage in tubeamps is known knowledge. It's cheap and it won't involve any
active components. A simple resistor or zobel network is all that is needed and it will prevent
damage.
An arduino however can be of great assistance in other areas. Two of them is to monitor cathod current
( voltage across a resistor) and shut down the amp while keeping the reason why and whick tube that is
the problem. Another function is to assist in powering up and down , preventing rapis off/on events,
and ensure proper timing at power on and off.
But acting on sudden flash is not where an arduino can be useful. (inho)
Hi Peter thank you for your reply. In my posts I suggested a transorb MOV or sparkgap will be required to handle the transient fault, the micro is to monitor the amp operation and apply a dummy load it necessary, in the case of a short circuit some amplifiers will red plate and damage tubes, a microprocessor based monitor can recognise this and take the appropriate action. A zobel network is also part of the solution for some fault conditions and often it is an essential part of the of the feedback-stability-loop.
A micro can also adjust tube bias, note tube aging, record tube hours, record amplifier run hours, detect RF or high frequency oscillations and lots more. Some $10 ESP32 boards have Wi-Fi and Bluetooth so remote control from a mobile phone is even possible.
Tubes and output transformers are so expensive these days a microcontroller based monitor-and-protect circuit is very cheap as a percentage of build cost. I was hoping the group may as a group design a cheap public domain device that enthusiasts not into micros and programming could easily use to protect their investment.
I just had a thought, a smoke detector module is <$5 if fitted it could record data to EEPROM and do a emergency shut down, maybe an optically coupled crowbar circuit could deliberately blow the mains fuse.
Ken