Hi all,
I'm putting together a series of laboratory experiments for classes I teach that involve several low-cost linear shaker motors - essentially high-excursion speaker motors with 8 ohm impedance.
Driving such motors is usually accomplished using laboratory grade power amplifiers with frequency response expending down to below 1 HZ. Those amps also run north of $3k for 100 watts RMS on a single channel.
My understanding is that DC amps should theoretically go down to 0Hz, which would enable a much lower cost solution. But nothing I can find on the market advertises a lower frequency response limit below 10 hz or so. Why is that? And is there any hope of locating an amp with a frequency response of say, 1Hz - 1 kHz?
I'm putting together a series of laboratory experiments for classes I teach that involve several low-cost linear shaker motors - essentially high-excursion speaker motors with 8 ohm impedance.
Driving such motors is usually accomplished using laboratory grade power amplifiers with frequency response expending down to below 1 HZ. Those amps also run north of $3k for 100 watts RMS on a single channel.
My understanding is that DC amps should theoretically go down to 0Hz, which would enable a much lower cost solution. But nothing I can find on the market advertises a lower frequency response limit below 10 hz or so. Why is that? And is there any hope of locating an amp with a frequency response of say, 1Hz - 1 kHz?
Servo is limiting factor. A servo Fc below 2-3hz is rare for an audio amp. I specialize in sub amps , I like <1hz!
DC blocking cap (in the design) will also roll off below a couple hz.
Below , my type amp will easily do 1hz , even with the safety of both a servo and AC input cap. I can watch my sub (pump) at 1-2 hz at nearly
full power. But not for long ... 1-2hz will overheat even a well vented , quality sub voice coil.
3K$ damn .... I could make a dozen 500W <1hz amps for that !!
Even without tweaking the circuit , 1hz is easy.
OS
DC blocking cap (in the design) will also roll off below a couple hz.
Below , my type amp will easily do 1hz , even with the safety of both a servo and AC input cap. I can watch my sub (pump) at 1-2 hz at nearly
full power. But not for long ... 1-2hz will overheat even a well vented , quality sub voice coil.
3K$ damn .... I could make a dozen 500W <1hz amps for that !!
Even without tweaking the circuit , 1hz is easy.
OS
Attachments
Wonder if those shaker motors have some sort of thermal dissipation setup ? I've made voice coils "glow" ... yikes !
There are Crown DC-300a used on ebay for <$200 . I can't link to the ad, ebay locked up my browser. But typing that string into bing jumped right to it.
May require an electrolytic cap replacment for full wattage and a trim potentiometer replacement for reliability, at 50 years of age. Heat sink & fan clean, too. No servo in those devices. Just 739 or 749 op amps.
May require an electrolytic cap replacment for full wattage and a trim potentiometer replacement for reliability, at 50 years of age. Heat sink & fan clean, too. No servo in those devices. Just 739 or 749 op amps.
DC coupled amplifiers go from DC to whatever their HF rating is.
Class D FM amplifiers work extremely well driving DC Servos.
Class D FM amplifiers work extremely well driving DC Servos.
If you want accurate LF from a large audio system, you need servo regulated, DC coupled amplification.
The issue is there are usually multiple coupling caps and HP filters in the entire signal path of most amplifiers and signal sources. This is often not necessary and allows for capacitor induced distortion to be introduced into your precious audio signal.
You do need at minimum a DC servo in the final amplification stage, which allows removal of most (if not all) AC coupled stages.
If you've ever heard a very good solid state amplified system, you'll be able to tell every tiny little nuance in the low end, but introducing even just a 10 hz HP will kill the phase accuracy, especially if the LF setup is sealed or critically dampened. Ported enclosures have their own issues and require careful filtering to avoid operation past the point of the enclosure unloading the LF driver, especially playing vinyl and other analog sources which can generate dangerous infrasonic noise.
So the effective argument here is - can you hear 10 hz on down? No, but you can hear the phase error of just one single stage of 10 hz HP. The problem worsens with multiple HP stages. The fewer, the better.
I've come to the conclusion that most people who argue against DC coupling have usually not heard really good sealed enclosure bass. Sorry, but ever since I removed as many HP stages as possible in my system, I've noticed a drastic change for the better in sound. The only thing left in my system with a HP is my phone stage, which uses a Lundahl coupling step up xformers. All of the capacitor coupled solutions I've tried so far didn't sound as good in the LF.
I
The issue is there are usually multiple coupling caps and HP filters in the entire signal path of most amplifiers and signal sources. This is often not necessary and allows for capacitor induced distortion to be introduced into your precious audio signal.
You do need at minimum a DC servo in the final amplification stage, which allows removal of most (if not all) AC coupled stages.
If you've ever heard a very good solid state amplified system, you'll be able to tell every tiny little nuance in the low end, but introducing even just a 10 hz HP will kill the phase accuracy, especially if the LF setup is sealed or critically dampened. Ported enclosures have their own issues and require careful filtering to avoid operation past the point of the enclosure unloading the LF driver, especially playing vinyl and other analog sources which can generate dangerous infrasonic noise.
So the effective argument here is - can you hear 10 hz on down? No, but you can hear the phase error of just one single stage of 10 hz HP. The problem worsens with multiple HP stages. The fewer, the better.
I've come to the conclusion that most people who argue against DC coupling have usually not heard really good sealed enclosure bass. Sorry, but ever since I removed as many HP stages as possible in my system, I've noticed a drastic change for the better in sound. The only thing left in my system with a HP is my phone stage, which uses a Lundahl coupling step up xformers. All of the capacitor coupled solutions I've tried so far didn't sound as good in the LF.
I
I recall visiting a test lab with a shaker table, circa 1970's. The "table" was perhaps 18 to 24 inches in diameter, maybe 3 feet high. Driven by a tube amplifier in a large rack cabinet--- locked unfortunately. I'm pretty sure the voice coil was a center-tapped affair and connected like the primary of push-pull audio transformer--- drivable to sub 1 Hz. I wish I could remember power, but somewhere in the 300 to 3kW class. A colleague concluded the table was capable of carrying his weight, sat on it and let it bounce him up and down. 😀
If you want to "roll your own" the attachment below from Analog Devices/Liner Tech features a 350W shaker table amp as an application.
All the advice in previous posts is good. You also may be able to find a used audio amp with sufficient power. As long as it features "direct-coupled" output drive and an available service manual, it's almost trivial to bypass a couple of capacitors and yield an amp with flat response to DC.
Have fun!
If you want to "roll your own" the attachment below from Analog Devices/Liner Tech features a 350W shaker table amp as an application.
All the advice in previous posts is good. You also may be able to find a used audio amp with sufficient power. As long as it features "direct-coupled" output drive and an available service manual, it's almost trivial to bypass a couple of capacitors and yield an amp with flat response to DC.
Have fun!
Modifying a Class AB audio amplifier for 1 Hz operation pushes the SOAR protection of the output transistors. You are looking at the DC curves
I'm ok with cutting under 5hz, but the problem still exists of how sensitive the protection circuit is and how it deals with the failsafe / fault condition ie disconnecting speakers via relay or crowbar circuit. I'm not a fan of relays if they sit closed constantly without cycling, so it needs to be integrated into the servo area, so that anything else that does make it through via a failed output device can't harm an LF driver.
I happen to know that some of rhe Columbia jazz remasters have alot of 2-3 hz junk riding on the PCM. I happen to catch it because my RME dac has DC detection built in, which flagged it immediately but didn't react, as I don't have the protection active on the dac side. Its clearly a software compression issue in their DAW.
I happen to know that some of rhe Columbia jazz remasters have alot of 2-3 hz junk riding on the PCM. I happen to catch it because my RME dac has DC detection built in, which flagged it immediately but didn't react, as I don't have the protection active on the dac side. Its clearly a software compression issue in their DAW.
Wow - some very helpful responses!
To clarify:
To clarify:
- I do not need true DC. A lower limit of 1Hz would be sufficient.
- Not a lot of wattage necessary. These little drivers are probably only rated for 10 or so watts.
- These are being packaged as a low-cost lab kit to be shared with other universities, so something off-the-shelf (some minor modifications are OK) would be ideal. 1HZ output is more important than phase-accuracy, THD, or any other audio quality considerations. That said, in going to look into the DC-300 and the attachments shared here for some higher end research applications.
If a typical amp with a split supply doesn't already run down to 1Hz, you just need to upgrade the input cap and the feedback decoupling cap a bit and you have a 1Hz amp. DC is also possible, but a servo has a cut off frequency where it sees low frequencies as offset. You might be better off with a offset trim-pot.
I ran a shake table for a few years, and we upgraded from an old 30KW cirlotron UDCO to a new 50KW IGBT class-D for a new 2" table. We never touched the potential power of the thing and I often considered smaller amps and tables. They had a long list of automatic shut-down features including a gradual ramp up that stops if the control accelerometer does read as expected.
I ran a shake table for a few years, and we upgraded from an old 30KW cirlotron UDCO to a new 50KW IGBT class-D for a new 2" table. We never touched the potential power of the thing and I often considered smaller amps and tables. They had a long list of automatic shut-down features including a gradual ramp up that stops if the control accelerometer does read as expected.
Drive motors with a motor driver perhaps? A class D amp is basically a motor driver circuit run at very high frequency with a low pass filter on the output. Using a linear amplifier to drive a motor is very inefficient, although it would be needed if audio-band noise is a problem - for a shaker motor I guess it depends on the bandwidth involved. Motor drivers are DC coupled always. Some class-D are DC coupled too.
It would help if you posted a link to the datasheet for the motors in question, as the winding resistance and inductance are important parameters you need to know.
As already mentioned above, a full-bridge class-D amplifier would go straight down to DC, like an SMPS.