Couldn't agree more. I suspect using some steel yoke type arrangements to steer the magnetic flux where you want it may be required.
Not much of a choice, the Reed type determines whether to use H31, H32, H33 or H34.
Sorry Jacco, I'm playing catch-up here. Your reference to the H31,H32 etc eludes me. Can you add more detail.
I use coarse attenuation jumps in a switched attenuator.
For fine adjustment, I find a continuous track vol pot much more convenient.
If very quiet, quiet, normal, loud, very loud, party level meets the listeners required range of SPL then a 6 step attenuator (-30dB to -0dB, or a bit wider maybe 10dB per step) would be good enough for me.
For fine adjustment, I find a continuous track vol pot much more convenient.
If very quiet, quiet, normal, loud, very loud, party level meets the listeners required range of SPL then a 6 step attenuator (-30dB to -0dB, or a bit wider maybe 10dB per step) would be good enough for me.
Standard rectangular magnets for Reed switches.
Width x Length ;
H31 : 0.06'' x 0.50''
H33 : 0.12'' x 0.75''
H32 : 0.19'' x 1.00''
H34 : 0.24'' x 1.00''
Reed switch diameter sets the magnet type, e.g. an H31 is used for a 0.07'' body diameter reed switch.
Okidoki.
For an example, the Reed switches in the Weiss attenuator are Hamlin's smallest, Miti-3(V).
Glass body of 1.78-1.80 mm diameter x 7mm length (0.07/0.71'' x 0.276'')
So the magnets used by Weiss are H31 (specified at 5mm/0.2'' distance from the Reed switch).
0.06'' height of an H31 magnet leaves so little to grab hold of, so to prevent them from dropping off, Weiss wrapped a transparent sleeve around the bar that holds the magnets.
The magnets are also 7/32'' longer than the glass body of the Reed switches, so Weiss had to position them higher, magnet top is almost flush with the solder tabs of the switches.
See piccy =>
For an example, the Reed switches in the Weiss attenuator are Hamlin's smallest, Miti-3(V).
Glass body of 1.78-1.80 mm diameter x 7mm length (0.07/0.71'' x 0.276'')
So the magnets used by Weiss are H31 (specified at 5mm/0.2'' distance from the Reed switch).
0.06'' height of an H31 magnet leaves so little to grab hold of, so to prevent them from dropping off, Weiss wrapped a transparent sleeve around the bar that holds the magnets.
The magnets are also 7/32'' longer than the glass body of the Reed switches, so Weiss had to position them higher, magnet top is almost flush with the solder tabs of the switches.
See piccy =>
Attachments
Thanks for all the feedback guys. Mercury wetted would be ideal, but as you mention they are way too expensive for the amounts required in this project. I've been looking at the various reed switch offerings and Hamlin seems like the way to go. Next step will be to choose a specific model and order enough to do a simple series attenuator first. Seems like the biggest challenge will be getting the mechanical aspects such as switch spacing and magnet choice just right to get a make-before-break action.
For the ladder attenuator version I'm thinking that stacking two SPST reed switches directly on top of each other for each step might work. I wonder how the magnet setup will have to be modified for that. This is where testing will be the only way to know what will and will not work. As was mentioned maybe some kind of steel yoke or surrounding each stacked reed pair with a small steel oval-tube sleeve would make them behave like a single DPST reed switch.
I'd like to have at least 60dB of total attenuation with 1dB per step. While doing professional audio work I tend to vary the monitor levels by quite a bit depending on what I'm doing. This is especially true when doing restoration and noise reduction work. When isolating noise sometimes I turn it way up then turn it back down to hear the result. So I'm all over the place which is why I'd like 1dB steps throughout the range. This is also why I want a tunable spring pressure on the detent. I would like a light amount of tension so that it can be turned up and down quickly.
For the ladder attenuator version I'm thinking that stacking two SPST reed switches directly on top of each other for each step might work. I wonder how the magnet setup will have to be modified for that. This is where testing will be the only way to know what will and will not work. As was mentioned maybe some kind of steel yoke or surrounding each stacked reed pair with a small steel oval-tube sleeve would make them behave like a single DPST reed switch.
I'd like to have at least 60dB of total attenuation with 1dB per step. While doing professional audio work I tend to vary the monitor levels by quite a bit depending on what I'm doing. This is especially true when doing restoration and noise reduction work. When isolating noise sometimes I turn it way up then turn it back down to hear the result. So I'm all over the place which is why I'd like 1dB steps throughout the range. This is also why I want a tunable spring pressure on the detent. I would like a light amount of tension so that it can be turned up and down quickly.
Decided to go with the Hamlin FLEX-14 series for the first prototype. This is the specific one: FLEX-14 15-20 Hamlin Inc | HE556-ND | DigiKey
They should be here sometime this coming week. These ones have long leads so they will be much easier to work with than the first ones I got. Looking forward to playing around. 🙂
They should be here sometime this coming week. These ones have long leads so they will be much easier to work with than the first ones I got. Looking forward to playing around. 🙂
Quick update. The new reed switches arrived last week. These are definitely going to be much easier to experiment with than the previous ones:
Here's the basic setup I'm going to use for testing:
This weekend I ordered a variety of tiny magnets to play with. Once they're here a new arm will be built to accommodate them.
Here's the basic setup I'm going to use for testing:
This weekend I ordered a variety of tiny magnets to play with. Once they're here a new arm will be built to accommodate them.
I did another round of testing using small neo cylinder magnets. In order to get a tighter spot of magnetic focus I tried a dual arm setup with a magnet on each side of the switch board. This worked much better at activating the dual stacked reed swtiches, but it was still inconsistent between adjacent positions.
I have no doubt that this would work for a single switch per position, series attenuator. However I think getting it to work 100% glitch free with a stacked pair of reed switches per position ladder attenuator will be a challenge. If someone still made a DPST reed switch it would be a different matter. Either way perhaps the biggest drawback to this whole approach is that it will be physically huge.
Now that the ELMA A47 is here I am shelving this reed switch attenuator project. For my passive monitor controller I am going to use a 10k quad A47 and just be done with it. Time to get my DAC and headphone amp projects done now.
I have no doubt that this would work for a single switch per position, series attenuator. However I think getting it to work 100% glitch free with a stacked pair of reed switches per position ladder attenuator will be a challenge. If someone still made a DPST reed switch it would be a different matter. Either way perhaps the biggest drawback to this whole approach is that it will be physically huge.
Now that the ELMA A47 is here I am shelving this reed switch attenuator project. For my passive monitor controller I am going to use a 10k quad A47 and just be done with it. Time to get my DAC and headphone amp projects done now.
+1.
I'm also a big reed fan, but physics wins out.
I thought of mounting the reeds "vertically" in a tight circle, with a bar magnet describing a smaller circle inside them. This would need some inter-reed screening since there is a lot of hysteresis in the reed.
A further thought-experiment would use a coil driven magnet whose flux could be reversed and varied to ensure adjacent reeds did not close.
But it never got beyond "just a thought!"
I'm also a big reed fan, but physics wins out.
I thought of mounting the reeds "vertically" in a tight circle, with a bar magnet describing a smaller circle inside them. This would need some inter-reed screening since there is a lot of hysteresis in the reed.
A further thought-experiment would use a coil driven magnet whose flux could be reversed and varied to ensure adjacent reeds did not close.
But it never got beyond "just a thought!"
I will start to programming on 64 steps Binary reed relay base attenuator soon but I worry about pop noise when switching. Who can tell me how to solve problem ? May be I will release some kits for DIY with reasonable price if I can solve problem on pop. I also had finished a 42 steps reed relay + VTC step attenuator with no pop but It cost too high. It need 24 reed relay per channel + multi-tap transformers. I try to developt a very high quality step attenuator with low cost to use in our integrated amplifiers.
for no pop:
no DC at the switches,
make before break.
as the least requirement.
Would a mute function that operates just before switching and releases just after switching be required?
no DC at the switches,
make before break.
as the least requirement.
Would a mute function that operates just before switching and releases just after switching be required?
A mute function could work, but in general, relay switched attenuators are hard to completely de-click, since they operate relatively slowly, and not completely predictably. With CMOS switches, folks have implemented 'zero crossing switching', which can work to some degree, but it's not so simple. For example, in the case of having to switch two or more channels, the zero crossings won't happen at the same time, so the switching will be randomly skewed according to the peculiarities of the signal.
Most folks who do relay switched attenuators just live with the clicks, and if you're careful, you can arrange the clicks to be somewhat small compared to the signal itself, e.g. with make before break switching, or any of a number of schemes to try to minimize disturbances.
Most folks who do relay switched attenuators just live with the clicks, and if you're careful, you can arrange the clicks to be somewhat small compared to the signal itself, e.g. with make before break switching, or any of a number of schemes to try to minimize disturbances.
Make before break is only workable in simple switching circuit but for Binary switch seem to be no way to do it. Due to it is not switch one switch to other. It is combination like binary digit. My design is not all reed relay. I use reed relay on signal path. Multiplexier for grounding. It take avantage of fast switching but higher resistor. I think it will not degrade sound quality but save much of space. I start to programming it today. I will start a new post when it sucessful.
Proper relay "timing" (different for increasing/decreasing attenuation)can solve all audible problems..I am using Lpad relay attenuators, 8relay, 0,5dB step. The only audible "clicking" is mechanical sound of relays.
Last edited:
"..since they operate relatively slowly, and not completely predictably..."
That is not true of miniature reed relays.
Switching time of about 1.5mS with very little variation. Software - or even hardware! - to do a mute, switch, open mute is pretty basic.
That is not true of miniature reed relays.
Switching time of about 1.5mS with very little variation. Software - or even hardware! - to do a mute, switch, open mute is pretty basic.
- Status
- Not open for further replies.