Hi Dinosauro,
Are you aware that most magic eyes tubes only last a few hundred hours at full light at the most? They are pretty much wasted as power indicators if you intend to have the amp powered up regularly.
Many radio collectors actually build switches into their vintage radios, so they can completely switch off the magic eye, heaters included, when operating the radio. Then only turning it on for photography sessions and on special occasions when they have visitors etc. This to save on their limited lifetime and the dwindling stock of spares available.
- Frank.
Are you aware that most magic eyes tubes only last a few hundred hours at full light at the most? They are pretty much wasted as power indicators if you intend to have the amp powered up regularly.
Many radio collectors actually build switches into their vintage radios, so they can completely switch off the magic eye, heaters included, when operating the radio. Then only turning it on for photography sessions and on special occasions when they have visitors etc. This to save on their limited lifetime and the dwindling stock of spares available.
- Frank.
Thus the qualifier 'most'. And so will any other type, which has the coating on the glass and not on a metal target. However EM84 and similar types are not nearly as popular for DIY due to looking more 'boring' to many people (different light color, simple rectangular stick/pinch pattern).
I am considering fitting EM84 or EM87 "magic eyes" to my home brew Mullard 5-20s. I would use a toggle switch to select between VU function and Bias set up for the output valves/tubes. Any comments?
I do not believe that the EM80 or other is accurate enough or stable enough to set bias using it as an indicator "Drift". You can drive it using a circuit "with some mods" I posted in a fun project from the link below:
http://www.diyaudio.com/forums/tubes-valves/177844-strip-board-project-completed-3.html
Some people like to drive with the HT from the anode of a pre-tube rather than audio signal. Your choice!
Regards
M. Gregg
Drift? I've never noticed any ... keep the voltages steady, and the indicator can be very stable.
Something like this might do the trick.
In general:
*) This configuration has the tube 'pinched' in idle and opening up as signal/current increases, Ie. the length of the indicator increases with input signal. The main problem with this setup is that the center part of the phosphor screen may suffer from 'burnout' rather quickly, reducing light output dramatically during idle. Yet on the other hand this also protected most of the 'active' area you actually need as much as possible.
*) I'd suggest making a test setup before knocking holes in your chassis etc. I know at least some people are a bit unhappy about the light output of magic eyes when run within specs. They are not daylight indicators.
*) Experiment with zener voltage to get the idle indication you want.
*) Voltages from a 5-20 into 8 ohm, 20W peak or so, is about right. With a 4 ohm load you may need a simple voltage doubler to get full deflection at peak signal.
*) R1 and R2 also serves to avoid popping noises when switching, so they shouldn't be left out completely. With R2 at 1Meg, you get about 75% deflection with a cathode voltage of 30V.
*) R5 is there to drop the anode voltage at pin 6 to max. 250V for max. longevity of the phosphor. The total current is about 1.8mA, so for instance with B+ at 440V, you'd use R5 ~ (440-250)/0.0018 = 105K, use 100K or 120K.
*) C2 smoothes out voltage fluctuations at pin 6 as the current from the EM84 changes with deflection. For a totally static display, for instance when displaying cathode current, you'd need this voltage stabilized somehow. C2 needs to be able to withstand the full B+ for starting up. Ie. use 450VW or higher for the 440V B+ I used as an example.
*) Grid on the left is pin 1, forgot to draw it in.
*) C1 to taste to get the delay response you like, or leave it out completely.
*) You could use an EB91/6AL5 dual diode for D1 if you can spare the heater current. Looks nice and would *slightly* improve linearity.
- Frank.
In general:
*) This configuration has the tube 'pinched' in idle and opening up as signal/current increases, Ie. the length of the indicator increases with input signal. The main problem with this setup is that the center part of the phosphor screen may suffer from 'burnout' rather quickly, reducing light output dramatically during idle. Yet on the other hand this also protected most of the 'active' area you actually need as much as possible.
*) I'd suggest making a test setup before knocking holes in your chassis etc. I know at least some people are a bit unhappy about the light output of magic eyes when run within specs. They are not daylight indicators.
*) Experiment with zener voltage to get the idle indication you want.
*) Voltages from a 5-20 into 8 ohm, 20W peak or so, is about right. With a 4 ohm load you may need a simple voltage doubler to get full deflection at peak signal.
*) R1 and R2 also serves to avoid popping noises when switching, so they shouldn't be left out completely. With R2 at 1Meg, you get about 75% deflection with a cathode voltage of 30V.
*) R5 is there to drop the anode voltage at pin 6 to max. 250V for max. longevity of the phosphor. The total current is about 1.8mA, so for instance with B+ at 440V, you'd use R5 ~ (440-250)/0.0018 = 105K, use 100K or 120K.
*) C2 smoothes out voltage fluctuations at pin 6 as the current from the EM84 changes with deflection. For a totally static display, for instance when displaying cathode current, you'd need this voltage stabilized somehow. C2 needs to be able to withstand the full B+ for starting up. Ie. use 450VW or higher for the 440V B+ I used as an example.
*) Grid on the left is pin 1, forgot to draw it in.
*) C1 to taste to get the delay response you like, or leave it out completely.
*) You could use an EB91/6AL5 dual diode for D1 if you can spare the heater current. Looks nice and would *slightly* improve linearity.
- Frank.
Attachments
There is an easy way around the problem with sagging supply voltages. Leave C2 completely out of the circuit, and instead buy five 47V/1W zener diodes. Connect these in series and tie between pin 6 and ground, zener cathodes to pin 6.
Recalculate R5 for an increased current of 0.5 mA, 2.3mA in total. The current through the EM84 decreases as the deflection increases, so this low current is fine.
For instance R5 ~ (440-5*47)/0.0023 = 89K, use 82K. This runs the EM84 at a supply voltage of 5*47-20 = 215V.
If you build two channels on one chassis, then you can share the zener diodes and R5 between a pair of EM84s. Then the current through R5 becomes 2*1.8+0.5mA = 4.1mA.
The total power dissipated by the zeners during startup is only 1W, shared between 5 diodes in total.
R5 needs to be some form of power resistor, considering current through and the voltage across it. 0.0041*(440-235) = 0.84. I'd probably use a 5W...
If running two EM84s in parallel, then each tube needs their own dummy anode resistor, the 470K R4.
Recalculate R5 for an increased current of 0.5 mA, 2.3mA in total. The current through the EM84 decreases as the deflection increases, so this low current is fine.
For instance R5 ~ (440-5*47)/0.0023 = 89K, use 82K. This runs the EM84 at a supply voltage of 5*47-20 = 215V.
If you build two channels on one chassis, then you can share the zener diodes and R5 between a pair of EM84s. Then the current through R5 becomes 2*1.8+0.5mA = 4.1mA.
The total power dissipated by the zeners during startup is only 1W, shared between 5 diodes in total.
R5 needs to be some form of power resistor, considering current through and the voltage across it. 0.0041*(440-235) = 0.84. I'd probably use a 5W...
If running two EM84s in parallel, then each tube needs their own dummy anode resistor, the 470K R4.
Thanks very much. That is similar to what I had in mind except for the toggle switch. I'm thinking of a 3 position switch to check both cathodes have roughly the right bias but, more importantly, DC balance. ( centre off can be used for VU ). I have a few EB91s, too.Something like this might do the trick.
In general:
*) This configuration has the tube 'pinched' in idle and opening up as signal/current increases, Ie. the length of the indicator increases with input signal. The main problem with this setup is that the center part of the phosphor screen may suffer from 'burnout' rather quickly, reducing light output dramatically during idle. Yet on the other hand this also protected most of the 'active' area you actually need as much as possible.
*) I'd suggest making a test setup before knocking holes in your chassis etc. I know at least some people are a bit unhappy about the light output of magic eyes when run within specs. They are not daylight indicators.
*) Experiment with zener voltage to get the idle indication you want.
*) Voltages from a 5-20 into 8 ohm, 20W peak or so, is about right. With a 4 ohm load you may need a simple voltage doubler to get full deflection at peak signal.
*) R1 and R2 also serves to avoid popping noises when switching, so they shouldn't be left out completely. With R2 at 1Meg, you get about 75% deflection with a cathode voltage of 30V.
*) R5 is there to drop the anode voltage at pin 6 to max. 250V for max. longevity of the phosphor. The total current is about 1.8mA, so for instance with B+ at 440V, you'd use R5 ~ (440-250)/0.0018 = 105K, use 100K or 120K.
*) C2 smoothes out voltage fluctuations at pin 6 as the current from the EM84 changes with deflection. For a totally static display, for instance when displaying cathode current, you'd need this voltage stabilized somehow. C2 needs to be able to withstand the full B+ for starting up. Ie. use 450VW or higher for the 440V B+ I used as an example.
*) Grid on the left is pin 1, forgot to draw it in.
*) C1 to taste to get the delay response you like, or leave it out completely.
*) You could use an EB91/6AL5 dual diode for D1 if you can spare the heater current. Looks nice and would *slightly* improve linearity.
- Frank.
Only my experience that when calibrated it seemed to go out of calibration after a few months. Also to get accuracy of less than 1V was difficult from just looking at the display. You had to wait for the tube to warm up and settle to get correct readings!
NB this was as a bias indicator VU seemed OK.
I'll have a look at Knarf's circuit.
Regards
M. Gregg
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.