Talk:High-Speed Uplink Packet Access
Some time ago this page was EUL, as the 3GPP suggested, and was mentioned in parentheses as HSUPA. Now this changed, to favor the name given by Nokia.
QPSK is a higher-order modulation with constant amplitude... adding section confusing tag. -- Taral 18:15, 16 May 2006 (UTC)
Comment about modulation
The modulation for HSUPA is BPSK instead of QPSK. We are talking here about Uplink (from UE to NodeB). The power required for an acceptable uplink transmission would be higher; this power increase will not only increase interference between UEs, but also the power output specifications in terms of SAR and performance of the battery, which would be reduced significantly (or batteries would become bigger). The higher bit rate is achieved by using smaller Spreading Factors and a combination of codes. Therefore, the modulation used for HSUPA is BPSK.
This is also "verified" if we analyze the numbers from the 3.8 Mchps (Mega Chips per second from the 5 Mhz carrier defined in R99) to the maximum defined bit rate of 5.76 Mbps.
To reach this speed, UE must support 2 x E-DPDCH codes with SF:2 (Spreading Factor) and 2 x E-DPDCH codes with SF:4. This leads to:
The 1 bit/chip comes from the modulation type. If modulation were QPSK this would mean 2 bit/chip and would double the physical bit rate.
As a reference from HSDPA, which uses 15 codes from a SF:16 and 16QAM (4 bit/chip) in the best conditions, the numbers look like:
(I hope this will help to clarify the topic. Comments are welcome.)
--Juanmre 23:47, 29 November 2006 (UTC)
Reply to Taral:
Thank you very much indeed for your comment. Here is what I would like to contribute to this discussion:
I thought, that this precisely was the point that the discussed paragraph made: QPSK is a constant envelope modulation, this is why it is used in UMTS and in this case in HSUPA. To help you understand the paragraph, here are the points, which it discusses:
1) A higher bit rate in the uplink is required. This can be achieved in different ways, one of which is to use a higher order modulation than the standard modulation in UMTS. The standard modulation in UMTS is QPSK, so using a modulation of a higher order than QPSK, such as 16QAM, which is used in HSDPA, could be one possible approach.
2) The modulations of a higher order than QPSK can be divided into:
- constant-envelope modulations;
- not-constant-envelope modulations.
The former are not very suitable for noisy channels as they require relatively high Eb/No for error-free demodulation. The latter, although performing better in noisy channels, have the disadvantage of requiring transmitters with higher power consumption, which is not desirable for a handheld mobile station.
3) Because of the above, after weighing the advantages and disadvantages of using a modulation of a higher order than QPSK, it was decided to keep QPSK in HSUPA and not to use a modulation of a higher order (as was done in HSDPA). Thus the increased bit rate in HSUPA is achieved by other means (such as a smaller spreading factor and a lower coding rate). Incidentally these methods too lead to reduced noise-tolerance and increased BER in the channel hence the introduction of HARQ to counteract this. At least, it was thought, QPSK allowed the existing low-cost, low-power-consumption transmitters to be used in HSUPA-capable mobile stations.
If you find the above a better worded and technically more accurate presentation of the subject, then would you like the discussed paragraph from the article to be replaced by a text similar to the above? If you think so, I will be prepared to edit the above and incorporate it in the article.
Additionally, I was thinking of some further clarifications, such as this 'side note' (see below). What would you think -- would you recommend this to be included in the article too? Will this help readers and will it reduce the scope for confusion?
A side note: Nothing comes out of nothing:
- If the bit rate is increased by using a robust and more noise-tolerant modulation (of higher order than QPSK), such as a non-constant-envelope modulation e.g. 16QAM, then the price is higher power consumption in the transmitter.
- If the bit rate is increased by using a constant-envelope modulation (of higher order than QPSK, such as M-point PSK with no zero crossing) then the price is a higher Eb/No required at the demodulator to produce the same BER. One possible way (but not the only one) to increase Eb/No at the receiver is to use higher transmitted power, which again leads to higher power consumption at the mobile station.
- If the bit rate is increased by using a smaller spreading factor and/or lower coding rate, then again the price will be the need of higher Eb/No to compensate for the reduced processing gain at the receiver. Again one way (but not the only one) to increase Eb/No at the receiver is to use higher transmitted power, which again leads to higher power consumption at the mobile station.
If the above three methods are considered for application in a single isolated channel, then they may appear largely equivalent in their result because the three will allow the bit rate to be increased at the cost of higher power-consumption at the transmitter. When these methods are considered for application in a system with multiple transmitters working simultaneously, we might see some advantages for the first method compared to the other two. This method leads to higher power-consumption at the transmitter but not higher transmitted power. The other two methods lead to increased transmitted power (and consequently higher power consumption too) in an effort to increase Eb/No at the receiver. However in a system with multiple transmitters working simultaneously it may not be possible to increase Eb/No at the receiver by increasing the power of each transmitter (increasing the power of a given transmitter will increase Eb at its corresponding receiver, but due to the interference from of all other transmitters, whose powers will be increased too, it is likely that No will also be increased and thus the ratio Eb/No may remain largely the same). For this reason it seemed initially attractive to apply to HSUPA the first method and to specify for it a modulation of a higher order than QPSK with non-constant envelope. As we said, this road was not taken in the HSUPA case due to some practical considerations, such as allowing simpler and more power-efficient transmitters to be used in the mobile stations. (This was the point of the discussed paragraph from the article). And because nothing comes out of nothing, the price for the increased bit rate in HSUPA is a reduced spreading factor and/or coding rate, which leads to higher Eb/No required at the demodulator, which, as we said, cannot be achieved by increasing the transmitted power of all transmitters in a system with multiple simultaneous transmitters. Hence HSUPA is expected to have a shorter range than the conventional UMTS uplink channels using the same modulation type, namely QPSK. (Well, by now the latter statement has become a kind of a platitude, apologies for this.)
Once again, thank you for your help and your comments.
Dual-Cell is just one of many extensions of the UMTS standard. It does not warrant a separate article. Dual-Cell HSDPA has already been merged into HSDPA. I suggest that DC-HSUPA be merged into HSUPA as well. --Drizzd (talk) 10:54, 17 June 2012 (UTC)