I am toying with the idea of throwing together a single-tube 1626 headphone amp using one of the Edcor GXSE transformers. My options are 5k:600 or 15k:600. My headphones are 300-ohm. So, I see the following options:
1) Use the 15k:600 with my headphones, giving me an effective 7.5k load.
2) Use the 5k:600 with a resistor in series with my headphones to get 600-ohm on the secondary.
Any bandwidth issues with option 1? Is this my better option? The 1626 should be fine with either load.
1) Use the 15k:600 with my headphones, giving me an effective 7.5k load.
2) Use the 5k:600 with a resistor in series with my headphones to get 600-ohm on the secondary.
Any bandwidth issues with option 1? Is this my better option? The 1626 should be fine with either load.
Last edited:
In principle, operating a transformer at half the design impedance will double the low end cut-off frequency - so there may be a lack of bass. However, whether or not it will be noticeable depends on the transformer design vis-a-vis circuit details. Bass may be limitted elsewhere as well. Also, the headphones you are using may have a resonance peak in the bass region which may counteract the bass loss in the transformer and give a better sound. So try it and see.
In general, putting resistance in series with speakers and headphones is not a good idea. It prevents the amplifier's output impedance from damping out speaker or headphone resonances, excessive treble response, and also leads to increased harmonic distortion at high volume levels. A triode offers some benefit in damping out resonances, distortion, and trebble rise, but it's not great. Don't make things worse by adding series resistors.
Keit
In general, putting resistance in series with speakers and headphones is not a good idea. It prevents the amplifier's output impedance from damping out speaker or headphone resonances, excessive treble response, and also leads to increased harmonic distortion at high volume levels. A triode offers some benefit in damping out resonances, distortion, and trebble rise, but it's not great. Don't make things worse by adding series resistors.
Keit
Last edited:
The results are depending on the output impedance of the circuit that is feeding the transformer.
According to the data sheet of 1626 it's plate resistance is 2500 ohms (with 25 mA plate current).
So it seems that 5k to 600 transformer will produce quite optimum load resistance.
Since the turns ratio of such transformer is 2,89 and the mu of 1626 is about 5 ( to 5.5.),
the net gain of the whole stage is about 5 dB, which seems quite good.
According to the data sheet of 1626 it's plate resistance is 2500 ohms (with 25 mA plate current).
So it seems that 5k to 600 transformer will produce quite optimum load resistance.
Since the turns ratio of such transformer is 2,89 and the mu of 1626 is about 5 ( to 5.5.),
the net gain of the whole stage is about 5 dB, which seems quite good.
Yes, the impedance matching (ie turns ratio) is good, but that wasn't NeedTube's question. He asked about bandwidth issues - as I said, potentially the bass end of the band will be lowered, but other factors may counteract it.
There's a typo in my first post - I meant halving the impedance will halve the bass cut-off frequency.
Keit
There's a typo in my first post - I meant halving the impedance will halve the bass cut-off frequency.
Keit
There is no such matter as transformer impedance.
There is the turns ratio that determines what impedance is reflected from secondary to primary ( and vice versa) when secondary is terminated with certain impedance.
Completely separate issue is the primary inductance that determines the low-end frequency response together with the source impedance of the device feeding the transformer.
So, if the load impedance is halved from 5k to 2.5k, but the transformer is driven by 2.5k
source impedance - as is the case with 1626 - the frequency response will remain unchanged.
I think this was NeedTube's question, and the answer is that there should be no bandwidth issues with 1626 driven 5k:600 transformer and 300 ohms headphones.
There is the turns ratio that determines what impedance is reflected from secondary to primary ( and vice versa) when secondary is terminated with certain impedance.
Completely separate issue is the primary inductance that determines the low-end frequency response together with the source impedance of the device feeding the transformer.
So, if the load impedance is halved from 5k to 2.5k, but the transformer is driven by 2.5k
source impedance - as is the case with 1626 - the frequency response will remain unchanged.
I think this was NeedTube's question, and the answer is that there should be no bandwidth issues with 1626 driven 5k:600 transformer and 300 ohms headphones.
And the 1626 will be fine driving the reflected 2500-ohms? Does this result in an increase in distortion? I still have yet to get a handle on transformer issues.
Compared to what ?
5k ?
I quicly drew a couple of loadlines to 1626 specs sheet.
The 2k5 loadline with Ua = 200 V. Ug = -22 V and Ia = 25 mA gave the max. output power of 400 mW. With same bias but with 5k load the output power is some 280 mW.
This tube seems to be designed for A2-operation and therefore the output power without going into grid current is relatively small, but should be well sufficient for headphone amplifier.
It is almost impossible to estimate from those loadlines which one gives better THD's.
There seems to be some confusion here. Probably not helped by my first post, typed in the middle of teh night after a very long day.
1. The low frequency cut-off is determined by the ratio of transformer inductance to the circuit impedance. The circuit impedance is essentaily the triode plate impedance in parallel with the reflected headphone impedance.
So, if you use a transformer at half the impedance it is designed for, the low frequency cut-off will be lower.
2. The high frequency cutoff of a transformer is determined by its internal leakage inductance - which arises because there is incomplete coupling between primary and secondary.
Thus, if you use a transformer at half the impedance it is designed for, the high frequency response will be worsened.
3. Whether or not you actually notice (1) and/or (2) depends on other factors. (2) is seldom a problem.
4. The impedance of headphones increases with frequency. This means that, unless they are driven from a low impedance, their response will increase with frequency. That's one reason why it is unwise to put resistors in series.
5. Loudspeakers and headphones generate distortion internally, as the voice coil moves through a magnetic field that isn't perfectly uniform. Driving them from a low inpedance reduces this distortion. This is because where the magnetic filed is strongest, the strongest back current is generated which flows through the driving impedance (ie the triode via the transformer).
This distortion has nothing to do with distortion generated by the triode itself, which principally arises from grid effects and teh three-halves rule of anode current - such distortion is usually not muchg affected by load impedance.
Keit
1. The low frequency cut-off is determined by the ratio of transformer inductance to the circuit impedance. The circuit impedance is essentaily the triode plate impedance in parallel with the reflected headphone impedance.
So, if you use a transformer at half the impedance it is designed for, the low frequency cut-off will be lower.
2. The high frequency cutoff of a transformer is determined by its internal leakage inductance - which arises because there is incomplete coupling between primary and secondary.
Thus, if you use a transformer at half the impedance it is designed for, the high frequency response will be worsened.
3. Whether or not you actually notice (1) and/or (2) depends on other factors. (2) is seldom a problem.
4. The impedance of headphones increases with frequency. This means that, unless they are driven from a low impedance, their response will increase with frequency. That's one reason why it is unwise to put resistors in series.
5. Loudspeakers and headphones generate distortion internally, as the voice coil moves through a magnetic field that isn't perfectly uniform. Driving them from a low inpedance reduces this distortion. This is because where the magnetic filed is strongest, the strongest back current is generated which flows through the driving impedance (ie the triode via the transformer).
This distortion has nothing to do with distortion generated by the triode itself, which principally arises from grid effects and teh three-halves rule of anode current - such distortion is usually not muchg affected by load impedance.
Keit
- Status
- Not open for further replies.