Speaker filters.. Reduce the hiss

[The Slab Bacon]

<snip> Much to my relief the 75A3 isn't in there.  It must be noiseless <snip>

Rob,
      Dont kid yourself, I have an 'A3, it SUCKS compared to some of my other receivers. It has the usual crappy Collins audio, and that noisy assed 6BA7 mixer make it one of my LEAST favorite receivers. The only thing that I like about it is the dial accuracy and stability of the Collins PTO. A slightly warmed over 390A blows it away!! Actually a good working SX-28 is a lot better receiver for AM.

                                                                The Slab Bacon

[WBear2GCR]

Geez Clark... calm down!!

A midrange will have a bandpass filter on it, yes?
That would be an inductor in series with cap going to the speaker. (at minimum)
I am quite confident that JBL never ever made a speaker or design where the impedance would drop to nil at some HF, even out of the audio band. So, no single caps to ground without something else in there to limit the impedance drop and limit the reactance.

Rob - the size of a speaker means absolutely nothing WRT the power of an audio amp.
Actually, large old 15" speakers are rather high sensitivity - meaning not much power in gets a pretty good power out.
So, it's fine as long as it has sufficient sensitivity. Typical speakers "back in the day" when it was a big deal to have a "record changer" in a fine "console" had to sound loud because the power available was typically <10watts on a good day going down hill!

Most modern "hi-fi" speakers are much lower in sensitivity - more on the 85db/1w/1m to about 90db/1w/1m and so need more power to reach equivalent loudness. Those older speakers and some newer ones are in the 100db/1w/1m class.

The newer guitar speaker that Clark has may have an aluminum dustcap and be designed to give a rising response or HF boost range, since it is a guitar speaker and not a hi-fi speaker... typically (seat-of-the-pants reckoning now...) a 10 ufd cap on an 8 ohm speaker would start rolling at about 2khz or so... maybe lower... which would mean that if the speaker was rising from around that range at 6dB/oct that this cap will flatten the response out to the point where the rise stops and then the whole thing will roll off neatly at a faster rate...  You can hear 10khz, but that doesn't tell you if the level is up or down, or if the response is or has been flattened.  If you want to try the inductor you need between a 1.0mh and 0.5mh - 0.65 being exactly at 2kHz @ 8ohms. Fwiw.

                    _-_-bear

[DMOD]

In decent communications speaker systems, there are at least some L and C components to shape bandwidth.

Take a look at the schematic of the SP-820

http://www.n6wk.com/kenwood/

Just scroll down to the SP-820 pdf file and check out the schematic there. I have changed the 40 F cap to 8 and 20 uFd switchables for better LF rolloff.

Regards,

Phil - AC0OB

[KM1H]

Steve, most all radios have audio hiss that starts at the detector. With a detector tube, 1st audio, phase inverter, and PP outputs that is a substantial string of noise generators with lots of gain. Scratch the phase inverter in some sets.

Pull the last IF tube and the noise is quite audible without having to crank up the volume. Especially with headphones.

Carl
KM1H

[W1RKW]

A hiss reducing or eliminating technique I have used on occasions is simply putting a rag over the speaker. 

[Pete, WA2CWA]

Now that I'm old, I don't hear hiss, although, now that I'm old, I'm happy I hear anything.

[Mike/W8BAC]

Go to jensentone.com and read the charts and specs. Most are 95 to 96 DB for 1 watt input.

95 to 96 DB of what? Are you saying you get 95 DB of sound with one watt? I want to see how that's measured.

[WBear2GCR]

The standard for loudspeaker (direct radiators, ie. cone speakers) is by the IEC.

For a given size speaker there is a standard size baffle, placed in an anechoic chamber. Then a 1 watt signal (2.83volts into an 8 ohm load - assuming this is an 8 ohm speaker) is fed to the speaker. A calibrated microphone is placed at a distance of ! meter.
The signal is swept sine wave. The resulting mic output is recorded on a log scale: dB vs. Frequency.

After that, it's a bit of a guessing game, and sometimes marketing hype, but for a wide range speaker the reference sensitivity is considered to be the region around 1 kHz.

The result is the "sensitivity" - sometimes called the "efficiency" but that is a misnomer - in dB. More precisely dB/1w/1m.
That is how most drivers are spec'd these days.

                    _-_-bear

[Mike/W8BAC]

Thanks Bear, That's interesting stuff. I'm having a hard time figuring out how, even at one meter, one watt of power can produce 95 DB at 1KHz. 95 DB is an awesome amount of sound. Exposure to 85 dB (sound pressure level - SPL) or higher can cause hearing loss over time. Maybe I'm missing something here.

[K5UJ]

It just seems to me that a 25 w. speaker would be taxing a little 2.5 w. audio amp and the amp would be struggling to make any noise at all with such a speaker.  There's a lot about audio amps and speakers and power I don't know.  my 12 w. p.p. tube amps can almost blow out my Polk CS1 speaker (120 watts).  I guess an amp can drive a x 10 power limit speaker okay.

Slabola -- I have been only running the A3 on 160 past night or two and condx have not been so good.  I don't use the stock 3 khz mech. filter -- I'm using a simple (and fairly wide) RC filter in the 455 KHz IF with the xtal filter out.   I'm contemplating a detector tap out to an external amp but wanted to hear the A3 direct to a decent speaker first.   Yeah, while I miss the big flywheel feel with a variable cap weighted shaft that you get with a NC300, I do like the accuracy and stability of the PTO.

Rob

[K5UJ]

Bear, looks like the 4 ohm jensen Mod 8-20 is the ticket:

http://jensentone.com/mod8-20.php

[WBear2GCR]

It just seems to me that a 25 w. speaker would be taxing a little 2.5 w. audio amp and the amp would be struggling to make any noise at all with such a speaker.

The way to understand this is that it is not a linear relationship 1 + 1 not= 2 in dB.

The "25 watt speaker" is a reference to the maximum power that the speaker can handle - usually this is a thermal limit, sometimes also the excursion limit at LF. There is no "taxing" at all... as long as you don't clip the amp, and don't exceed the thermal and excursion limits of a speaker, anything inside that range is ok fine!

You have to understand that most speakers are less than 1% efficient in terms of turning electricity into sound!! It is mostly turned into heat.

The 1 watt SPL rating is the key to understanding this. At one watt of input power most speakers will produce an adequate and "normal level" sound output. It takes DOUBLE the power to increase the percieved output by "two just noticeable differences" that is 3dB. So if the speaker is typical and produces 90dB/1w/1m it will take 8 watts to get to 99dB. But sound falls off with distance... so unless you are sitting nearfield, you will be getting less and less SPL the farther away you are... (a big problem with PA/SR systems).

 There's a lot about audio amps and speakers and power I don't know.  my 12 w. p.p. tube amps can almost blow out my Polk CS1 speaker (120 watts).  I guess an amp can drive a x 10 power limit speaker okay.


Rob

[WBear2GCR]

Thanks Bear, That's interesting stuff. I'm having a hard time figuring out how, even at one meter, one watt of power can produce 95 DB at 1KHz. 95 DB is an awesome amount of sound. Exposure to 85 dB (sound pressure level - SPL) or higher can cause hearing loss over time. Maybe I'm missing something here.

Actually this is a partial truth - otherwise we'd all be stone deaf long ago... maybe government OSHA stuff?

Long exposures at highish levels can cause long term damage
SHORT exposures at VERY HIGH levels will cause damage.

95dB at your ears is a somewhat loud sound - far far less than you hear in a typical concert. That's about the level that you'd listen to Led Zep if you like it just slightly loudish... not crushing...

                   _-_-bear

[Steve - WB3HUZ]

Yup. But that ain't band noise. It's crappy design. Either way, a lowpass filter to remove the hiss will also lowpass filter the desired audio too. As you know TANSTAAFL.

Steve, most all radios have audio hiss that starts at the detector. With a detector tube, 1st audio, phase inverter, and PP outputs that is a substantial string of noise generators with lots of gain. Scratch the phase inverter in some sets.

Pull the last IF tube and the noise is quite audible without having to crank up the volume. Especially with headphones.

Carl
KM1H

[WBear2GCR]

Well actually a LPF set between 5-7kHz ought to be quite ok... especially since most transmitted bandwidth falls within that range. Most, not all. Most of the hiss is heard as HF noise, so that ought to reduce a lot of it. If you still have noise after that, then I'd be looking into the receiver a bit more, or doing something to get more signal into the front end??

                       _-_-bear

[WBear2GCR]

Bear, looks like the 4 ohm jensen Mod 8-20 is the ticket:

http://jensentone.com/mod8-20.php

I'd not get a 4 ohm speaker to use with boatanchor gear. 8 ohms or higher is better.

The speaker you mention here is an 8" driver.
For me that is smallish.
However, I do use a JBL D-208 (iirc the number)in a sealed cabinet with my TS 440. It has wonderful sound.
That is an 8" driver. Vintage.
Look it up.

In the 15" size the JBL D-120 or K-140 is worthy too...

I looked at the Jensen guitar speakers, I'd not want to use any of them for listening to AM.
Well, maybe if I could have one of each and try them all??

One might consider other drivers. Parts Express sells a ton of drivers as does MCM. There are other places too.
Eminence has a wide range of stuff that is high sensitivity too.

But one of the best solutions is to find a local flea market or craiglist listing for old console "hifi" and "borrow" the whole inside baffle, drivers and all for the job! Some are quite nice, with a decent paper cone midrange and tweeter. Usually the tweeter can be dispensed with or padded down, but those old paper cone woofers and midranges sound great on voices.

                    _-_-bear

[KM1H]

Yup. But that ain't band noise. It's crappy design. Either way, a lowpass filter to remove the hiss will also lowpass filter the desired audio too. As you know TANSTAAFL.

Steve, most all radios have audio hiss that starts at the detector. With a detector tube, 1st audio, phase inverter, and PP outputs that is a substantial string of noise generators with lots of gain. Scratch the phase inverter in some sets.

Pull the last IF tube and the noise is quite audible without having to crank up the volume. Especially with headphones.

Carl
KM1H

Aint nothing wrong with the design, just components available at the time.  Im sure that replacing all audio stage resistors with metal film and adapting the lowest noise tubes that were available much later would help a lot.

However a simple LPF is a lot less work.

In the 80's a bunch of us in the YCCC and other contest clubs bought a special design Beyer headphone built for communications audio. Its fantastic for no fatigue long operating sessions.  Heil had their own flyweight version that is OK but the Beyer is mo betta. A lot pricer also at $250 back then.

Carl

[K5UJ]

Thanks Bear; lots of options.  We had an old console hifi that belonged to my parents.  It was a big piece of furniture with a tube amp, AM/FM stereo tuner, and Girard turntable.  One of those wide low cabinet things with a lid that lifted up in the middle for the turntable and tuner.  Speakers on each end.   I sure wish I had that now but we let it go a couple of years ago with some other things. 

[Ian VK3KRI]

Thanks Bear, That's interesting stuff. I'm having a hard time figuring out how, even at one meter, one watt of power can produce 95 DB at 1KHz. 95 DB is an awesome amount of sound. Exposure to 85 dB (sound pressure level - SPL) or higher can cause hearing loss over time. Maybe I'm missing something here.

Some years ago I had a couple of  18" metal folded horn PA speakers using a compression driver. Awful, restricted, lumpy  frequency response , but because of the horn loading, super efficient at the frequencies it actually worked at.  A pocket transistor radio producing maybe 100mW of audio, was mind numbingly loud thru that speaker, especially when you compared it against the pitiful internal 1 1/2" thing.  One Watt of audio coupled to the air at 100% efficiency would be astonishingly loud.

Also too , don't forget that when we listen to a audio amp, were normally listening to music or talk. The peak to average ratio is quite high, so your 100W amp might be putting out an average of 1 watt possibly less , even thought the peaks are close to 100.


And back to the topic, almost, Arn't the caps across the primaries of output transformers there to form a "build out" network with the leakage inductance of the transformer that extends the frequency response  at expense of an ultimately faster roll off  past the cutoff point, or am I confused, or just deaf - again.
                                                                                    Ian VK3KRI

[ke7trp]

Yep. 1 watt 1 meter.  96 is pretty good. But not loud by any means.  Years ago, I competed in IASCA car audio competitions. My car was built for sound quality. But it would still produce 129 DB.

I have now toggled the cap in an out 100 times. Listening to Don last night.  It has no effect on dons audio. NONE at 10 UF.  Just the hiss of the SP600 is gone. What a treat the cap is. I just love it. 

The SX28 does not need the cap. It has a built in tone control. Hmm. I wonder how that works?

The rating of the speaker is the MAX power that is supported.  Its really just a general rating. Everyone knows that the Box the speaker is in has more impact on the power handling then anything else. The Xmax is very important

A Lower power rated speaker does not play louder then a higher power rating one most of the time. Unless, its motor assembly has been made very stiff for very high power or LOW Freq use. Then, You will have a lower SPL rating.

I have the 8 ohm model 10-30 MOD. Its just great considering its just $34.  I would like the original style Alnico which is still made. But its around $100. Its just not worth the extra money to me for ham radio reciever.

Today, I installed a rack shelf and hooked a $2000 dollar monitor speaker up to the SX28.  Why?  Cause I got both the monitors cheap for resale. Then realised, One had a blown tweeter and the other had a blown woofer. So I made one good speaker out of the set. I cant sell it. Nobody wants one speaker.. LOL.  The SX 28 sounds like FM. Maybe better playing oldies in the shack.  Just shocked out how good this thing sounds! Its up there with the SX62 which is clearly the king of the HiFi receivers...



C

[K5UJ]

I heard somewhere that the SX25 is a sleeper and sounds pretty good too, something supposedly that is not well known.

Okay 8 ohms it is.  As is often the case, the audio xformers in my two amps have taps for 4, 8 and 16 ohms.  The speakers I use are 8 ohms and the 4 ohm tap sounded the best. 

[K5UJ]

SX62 guys--go for it:  http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=300402979156

[KM1H]

The SX-16, 17 and 25 all use PP 6F6's which arent the best of tubes in pentode mode but they sound OK on the ham bands. I have all 3 here.

The SX-42 and the various SX-62's all share the same audio section. Hallicrafters claims its "flat" out to 15 kc.

I have a pair of SX-62A's all rebuilt. One drives a R-42 bass reflex speaker and sits in the bedroom. The other has the audio section disabled (called pull the tubes) and the phono-in is now audio out. That drives a mid 50's 4 speaker (2 each 12" and 4") RCA AM/FM/Phono console audio deck which has much better control (more knobs to twist) of the audio response. That is in the LR and can rattle the windows. That and the Scott 800B (in the DR) mentioned below get the most comment here at family gatherings

The next on the rebuild list is a SP-400. I had a BC-779 in the 60's and the triode connected 6F6's plus the triode strapped 6F6 driver sounded super into an old 15" coaxial Jensen which probably came from a scrapped Scott 800B. (Anyway its identical to the one in the 800B I have now) The problem with the 779 is that it never stopped drifting.

Carl
KM1H

[WD5JKO]

OK guys, you got me to try this. I took my Icom R75 SWL receiver and tried 10uf, then 20 uf across the speaker. NO CHANGE! The R-75 on AM with a 6 Khz IF filter also has a 3 Khz multi-pole audio LPF with op-amps.

So thinking, the solid state audio probably had a very low source impedance, much different than a Beam Power tube amp without NFB. So I added a 10 ohm resistor in series with the receiver so that the cap works against a series resistor. Had to crank the audio gain a bit, but the headroom was there. Thinking that a 10uf non polar cap has about 10 ohms Xc at 1600 Hz, I thought that would be too much. Adding the cap makes a difference, but barely to my ears. That is until I tune off away from signals in the presence of impulse type QRN. HUGE CHANGE! Now that background is not nearly as irritating.

I will have to play with this idea for a while.

Jim
WD5JKO

[WBear2GCR]

Hang on a second!

What sort of speaker are we talking about??

... a few points, the cap's effect is WRT the impedance of the speaker - it has little to do with the output impedance of the power amplifier, tube or solid state.

If the output of the receiver already has a 3kHz LPF engaged, the addition of a cap of the value used would have little effect.

If the series resistor and the impedance of the cap are identical, then you've got a 6dB pad (at that frequency)... and with the series resistor a pad made with the impedance of the speaker itself - same idea if it was a 10 ohm resistor and 10 ohm speaker you'd again have a 6dB drop in level...

Again, I would strongly suggest the use of a small inductor over a cap in parallel. If you need more rapid roll off, then add a cap in parallel to the speaker to get a 12db/oct rolloff rate... you can shelve the rolloff with a resistor or two (depending on if you use one component or two).

Just do a quick search online for "crossover calculator" and you'll find any number of sites that allow you to plunk in various compenents and impedances and get back the nominal turnover frequency. A search for "crossover simulator" might bring up some freeware and online utilities that include graphics as well. LTspice will do this as well, but it is more complex to learn how to run...

                _-_-bear